add fcn and voc

chapter_computer-vision/fcn.md

-# 语义分割
+# 全连接卷积网络：FCN
 
-我们已经学习了如何识别图片里面的主要物体，和找出里面物体的边框。语义分割则在之上更进一步，它对每个像素预测它是否只是背景，还是属于哪个我们感兴趣的物体。
+在图片分类里，我们通过卷积层和池化层逐渐减少图片高宽最终得到跟预测类别数长的向量。例如用于ImageNet分类的ResNet 18里，我们将高宽为224的输入图片首先减少到高宽7，然后使用全局池化层得到512维输出，最后使用全连接层输出长为1000的预测向量。
 
-![Semantic Segmentation](../img/segmentation.svg)
+但在语义分割里，我们需要对每个像素预测类别，也就是需要输出形状需要是$1000\times 224\times 224$。如果仍然使用全链接层作为输出，那么这一层权重将多达数百GB。本小节我们将介绍利用卷积神经网络解决语义分割的一个开创性工作之一：全链接卷积网络（fully convolutional network，简称FCN）[1]。FCN里将最后的全连接层修改称转置卷积层（transposed convolution）来得到所需大小的输出。
 
-可以看到，跟物体检测相比，语义分割预测的边框更加精细。
-
-也许大家还听到过计算机视觉里的一个常用任务：图片分割。它跟语义分割类似，将每个像素划分的某个类。但它跟语义分割不同的时候，图片分割不需要预测每个类具体对应哪个物体。因此图片分割经常只需要利用像素之间的相似度即可，而语义分割则需要详细的类别标号。这也是为什么称其为**语义**的原因。
-
-
-本章我们将介绍利用卷积神经网络解决语义分割的一个开创性工作之一：[全链接卷积网络](https://arxiv.org/abs/1411.4038)。在此之前我们先了解用来做语义分割的数据。
-
-## 数据集
-
-[VOC2012](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/)是一个常用的语义分割数据集。输入图片跟之前的数据集类似，但标注也是保存称相应大小的图片来方便查看。下面代码下载这个数据集并解压。注意到压缩包大小是2GB，可以预先下好放置在`data_root`下。
-
-```{.python .input  n=1}
+```{.python .input  n=2}
 %matplotlib inline
 import sys
 sys.path.append('..')
 import gluonbook as gb
-import tarfile
-from mxnet import nd, image, gluon, init
+from mxnet import gluon, init, nd, image
 from mxnet.gluon import nn
 import numpy as np
 ```
 
-```{.python .input  n=2}
-data_root = '../data'
-voc_root = data_root + '/VOCdevkit/VOC2012'
-url = ('http://host.robots.ox.ac.uk/pascal/VOC/voc2012'
-       '/VOCtrainval_11-May-2012.tar')
-sha1 = '4e443f8a2eca6b1dac8a6c57641b67dd40621a49'
-
-fname = gluon.utils.download(url, data_root, sha1_hash=sha1)
+## 转置卷积层
 
-with tarfile.open(fname, 'r') as f:
-    f.extractall(data_root)
-```
+假设$f$是一个卷积层，给定输入$x$，我们可以计算前向输出$y=f(x)$。在反向求导$z=\frac{\partial\, f(y)}{\partial\,x}$时，我们知道$z$会得到跟$x$一样形状的输出。因为卷积运算的导数的导数是自己本身，我们可以合法定义转置卷积层，记为$g$，为交互了前向和反向求导函数的卷积层。也就是$z=g(y)$。
 
-下面定义函数将训练图片和标注按序读进内存。
+下面我们构造一个卷积层并打印其输出形状。
 
 ```{.python .input  n=3}
-def read_images(root=voc_root, train=True):
-    txt_fname = '%s/ImageSets/Segmentation/%s'%(
-        root, 'train.txt' if train else 'val.txt')
-    with open(txt_fname, 'r') as f:
-        images = f.read().split()
-    data, label = [None] * len(images), [None] * len(images)
-    for i, fname in enumerate(images):
-        data[i] = image.imread('%s/JPEGImages/%s.jpg'%(root, fname))
-        label[i] = image.imread('%s/SegmentationClass/%s.png'%(root, fname))
-    return data, label
-
-train_images, train_labels = read_images()
-```
-
-我们画出前面三张图片和它们对应的标号。在标号中，白色代表边框黑色代表背景，其他不同的颜色对应不同物体。
-
-```{.python .input  n=4}
-n = 5
-imgs = train_images[0:n] + train_labels[0:n]
-gb.show_images(imgs, 2, n);
-```
-
-同时注意到图片的宽度基本是500，但高度各不一样。为了能将多张图片合并成一个批量来加速计算，我们需要输入图片都是同样的大小。之前我们通过`imresize`来将他们调整成同样的大小。但在语义分割里，我们需要对标注做同样的变化来达到像素级别的匹配。但调整大小将改变像素颜色，使得再将它们映射到物体类别变得困难。
-
-这里我们仅仅使用剪切来解决这个问题。就是说对于输入图片，我们随机剪切出一个固定大小的区域，然后对标号图片做同样位置的剪切。
-
-```{.python .input  n=5}
-def rand_crop(data, label, height, width):
-    data, rect = image.random_crop(data, (width, height))
-    label = image.fixed_crop(label, *rect)
-    return data, label
-
-imgs = []
-for _ in range(n):
-    imgs += rand_crop(train_images[0], train_labels[0], 200, 300)
-gb.show_images(imgs[::2]+imgs[1::2], 2, n);
-```
-
-接下来我们列出每个物体和背景对应的RGB值
-
-```{.python .input  n=6}
-classes = ['background','aeroplane','bicycle','bird','boat',
-           'bottle','bus','car','cat','chair','cow','diningtable',
-           'dog','horse','motorbike','person','potted plant',
-           'sheep','sofa','train','tv/monitor']
-# RGB color for each class
-colormap = [[0,0,0],[128,0,0],[0,128,0], [128,128,0], [0,0,128],
-            [128,0,128],[0,128,128],[128,128,128],[64,0,0],[192,0,0],
-            [64,128,0],[192,128,0],[64,0,128],[192,0,128],
-            [64,128,128],[192,128,128],[0,64,0],[128,64,0],
-            [0,192,0],[128,192,0],[0,64,128]]
-
-len(classes), len(colormap)
-```
-
-这样给定一个标号图片，我们就可以将每个像素对应的物体标号找出来。
-
-```{.python .input  n=7}
-import numpy as np
-
-cm2lbl = nd.zeros(256**3)
-for i,cm in enumerate(colormap):
-    cm2lbl[(cm[0]*256+cm[1])*256+cm[2]] = i
-
-def image2label(im):
-    data = im.astype('int32').asnumpy()
-    idx = (data[:,:,0]*256+data[:,:,1])*256+data[:,:,2]
-    return cm2lbl[idx]
-```
-
-可以看到第一张训练图片的标号里面属于飞机的像素被标记成了1.
-
-```{.python .input  n=8}
-y = image2label(train_labels[0])
-y[105:115, 130:140]
-```
-
-现在我们可以定义数据读取了。每一次我们将图片和标注随机剪切到要求的形状，并将标注里每个像素转成对应的标号。简单起见我们将小于要求大小的图片全部过滤掉了。
-
-```{.python .input  n=9}
-rgb_mean = nd.array([0.485, 0.456, 0.406])
-rgb_std = nd.array([0.229, 0.224, 0.225])
-
-def normalize_image(data):
-    return (data.astype('float32') / 255 - rgb_mean) / rgb_std
-
-class VOCSegDataset(gluon.data.Dataset):
-
-    def _filter(self, images):
-        return [im for im in images if (
-            im.shape[0] >= self.crop_size[0] and
-            im.shape[1] >= self.crop_size[1])]
-
-    def __init__(self, train, crop_size):
-        self.crop_size = crop_size
-        data, label = read_images(train=train)
-        data = self._filter(data)
-        self.data = [normalize_image(im) for im in data]
-        self.label = self._filter(label)
-        print('Read '+str(len(self.data))+' examples')
-
-    def __getitem__(self, idx):
-        data, label = rand_crop(
-            self.data[idx], self.label[idx],
-            *self.crop_size)
-        data = data.transpose((2,0,1))
-        label = image2label(label)
-        return data, label
-
-    def __len__(self):
-        return len(self.data)
-```
-
-我们采用$320\times 480$的大小用来训练，注意到这个比前面我们使用的$224\times 224$要大上很多。但是同样我们将长宽都定义成了32的整数倍。
-
-```{.python .input  n=10}
-# height x width
-input_shape = (320, 480)
-voc_train = VOCSegDataset(True, input_shape)
-voc_test = VOCSegDataset(False, input_shape)
-```
-
-最后定义批量读取。可以看到跟之前的不同是批量标号不再是一个向量，而是一个三维数组。
-
-```{.python .input  n=11}
-batch_size = 64
-train_data = gluon.data.DataLoader(
-    voc_train, batch_size, shuffle=True,last_batch='discard')
-test_data = gluon.data.DataLoader(
-    voc_test, batch_size,last_batch='discard')
+conv = nn.Conv2D(10, kernel_size=4, padding=1, strides=2)
+conv.initialize()
 
-for data, label in train_data:
-    print(data.shape)
-    print(label.shape)
-    break
+x = nd.random.uniform(shape=(1,3,64,64))
+y = conv(x)
+y.shape
 ```
 
-## 全连接卷积网络
-
-在数据的处理过程我们看到语义分割跟前面介绍的应用的主要区别在于，预测的标号不再是一个或者几个数字，而是每个像素都需要有标号。在卷积神经网络里，我们通过卷积层和池化层逐渐减少数据长宽但同时增加通道数。例如ResNet18里，我们先将输入长宽减少32倍，由$3\times 224\times 224$的图片转成$512\times 7 \times 7$的输出，应该全局池化层变成$512$长向量，然后最后通过全链接层转成一个长度为$n$的输出向量，这里$n$是类数，既`num_classes`。但在这里，对于输出为$3\times 320 \times 480$的图片，我们需要输出是$n \times 320 \times 480$，就是每个输入像素都需要预测一个长度为$n$的向量。
-
-全连接卷积网络（FCN）的提出是基于这样一个观察。假设$f$是一个卷积层，而且$y=f(x)$。那么在反传求导时，$\partial f(y)$会返回一个跟$x$一样形状的输出。卷积是一个对偶函数，就是$\partial^2 f = f$。那么如果我们想得到跟输入一样的输入，那么定义$g = \partial f$，这样$g(f(x))$就能达到我们想要的。
+使用用样的卷积窗、填充和步幅的转置卷积层，我们可以得到跟`x`一样的输出。
 
-具体来说，我们定义一个卷积转置层（transposed convolutional, 也经常被错误的叫做deconvolutions），它就是将卷积层的`forward`和`backward`函数兑换。
-
-下面例子里我们看到使用同样的参数，除了替换输入和输出通道数外，`Conv2DTranspose`可以将`nn.Conv2D`的输出还原其输入大小。
-
-```{.python .input  n=12}
-conv = nn.Conv2D(10, kernel_size=4, padding=1, strides=2)
+```{.python .input  n=4}
 conv_trans = nn.Conv2DTranspose(3, kernel_size=4, padding=1, strides=2)
-
-conv.initialize()
 conv_trans.initialize()
-
-x = nd.random.uniform(shape=(1,3,64,64))
-y = conv(x)
-print('Input:', x.shape)
-print('After conv:', y.shape)
-print('After transposed conv', conv_trans(y).shape)
+conv_trans(y).shape
 ```
 
-另外一点要注意的是，在最后的卷积层我们同样使用平化层（`nn.Flattern`）或者（全局）池化层来使得方便使用之后的全连接层作为输出。但是这样会损害空间信息，而这个对语义分割很重要。一个解决办法是去掉不需要的池化层，并将全连接层替换成$1\times 1$卷基层。
+简单来说，卷积层通常使得输入高宽变小，而转置卷积层则一般用来将高宽增大。
 
-所以给定一个卷积网络，FCN主要做下面的改动
+## FCN模型
 
-- 替换全连接层成$1\times 1$卷基
-- 去掉过于损失空间信息的池化层，例如全局池化
-- 最后接上卷积转置层来得到需要大小的输出
-- 为了训练更快，通常权重会初始化称预先训练好的权重
+FCN的核心思想是将一个卷积网络的最后全连接输出层替换成转置卷积层来获取对每个输入像素的预测。具体来说，它去掉了过于损失空间信息的全局池化层，并将最后的全链接层替换成输出通道是原全连接层输出大小的$1\times 1$卷积层，最后接上卷积转置层来得到需要形状的输出。
 
-![FCN](../img/fcn.svg)
+![FCN模型。](../img/fcn.svg)
 
-下面我们基于Resnet18来创建FCN。首先我们下载一个预先训练好的模型。
+下面我们基于ResNet 18来创建FCN。首先我们下载一个预先训练好的模型，并打印其最后的数个神经层。
 
-```{.python .input  n=13}
+```{.python .input  n=3}
 pretrained_net = gluon.model_zoo.vision.resnet18_v2(pretrained=True)
 
 (pretrained_net.features[-4:], pretrained_net.output)
 ```
 
-我们看到`feature`模块最后两层是`GlobalAvgPool2D`和`Flatten`，都是我们不需要的。所以我们定义一个新的网络，它复制除了最后两层的`features`模块的权重。
+可以看到`feature`模块最后两层是`GlobalAvgPool2D`和`Flatten`，在FCN里均不需要，`output`模块里的全链接层也需要舍去。下面我们定义一个新的网络，它复制除了`feature`里除去最后两层的所有神经层以及权重。
 
-```{.python .input  n=14}
+```{.python .input  n=4}
 net = nn.HybridSequential()
 for layer in pretrained_net.features[:-2]:
     net.add(layer)
+```
 
-x = nd.random.uniform(shape=(1,3,*input_shape))
-print('Input:', x.shape)
-print('Output:', net(x).shape)
+给定高宽为224的输入，`net`的输出将输入高宽减少了32倍。
+
+```{.python .input}
+x = nd.random.uniform(shape=(1,3,224,224))
+net(x).shape
 ```
 
-然后接上一个通道数等于类数的$1\times 1$卷积层。注意到`net`已经将输入长宽减少了32倍。那么我们需要接入一个`strides=32`的卷积转置层。我们使用一个比`stides`大两倍的`kernel`，然后补上适当的填充。
+为了是的输出跟输入有同样的高宽，我们构建一个步幅为32的转置卷积层，卷积核的窗口高宽设置成步幅的2倍，并补充适当的填充。在转置卷积层之前，我们加上$1\times 1$卷积层来将通道数从512降到标注类别数，对Pascal VOC数据集来说是21。
 
-```{.python .input  n=15}
-num_classes = len(classes)
+```{.python .input  n=5}
+num_classes = 21
 
 net.add(
     nn.Conv2D(num_classes, kernel_size=1),
@@ -245,11 +79,11 @@ net.add(
 )
 ```
 
-## 训练
+## 模型初始化
 
-训练的时候我们需要初始化新添加的两层。我们可以随机初始化，但实际中发现将卷积转置层初始化成双线性差值函数可以使得训练更容易。
+`net`中的最后两层需要对权重进行初始化，通常我们会使用随机初始化。但新加入的转置卷积层的功能有些类似于将输入调整到更大的尺寸。在图片处理里面，我们可以通过有适当卷积核的卷积运算符来完成这个操作。常用的包括双线性差值核，下面函数构造核权重。
 
-```{.python .input  n=16}
+```{.python .input  n=6}
 def bilinear_kernel(in_channels, out_channels, kernel_size):
     factor = (kernel_size + 1) // 2
     if kernel_size % 2 == 1:
@@ -264,84 +98,81 @@ def bilinear_kernel(in_channels, out_channels, kernel_size):
         dtype='float32')
     weight[range(in_channels), range(out_channels), :, :] = filt
     return nd.array(weight)
-
 ```
 
-下面代码演示这样的初始化等价于对图片进行双线性差值放大。
-
-```{.python .input  n=17}
-x = train_images[0]
-print('Input', x.shape)
-x = x.astype('float32').transpose((2,0,1)).expand_dims(axis=0)/255
+接下来我们构造一个步幅为2的转置卷积层，将其权重初始化成双线性差值核。
 
-conv_trans = nn.Conv2DTranspose(
-    3, in_channels=3, kernel_size=8, padding=2, strides=4)
-conv_trans.initialize()
-conv_trans(x)
-conv_trans.weight.set_data(bilinear_kernel(3, 3, 8))
+```{.python .input  n=7}
+conv_trans = nn.Conv2DTranspose(3, kernel_size=4, padding=1, strides=2)  
+conv_trans.initialize(init.Constant(bilinear_kernel(3, 3, 4)))
+```
 
+可以看到这个转置卷积层的前向函数的效果是将输入图片高宽扩大2倍。
 
+```{.python .input}
+img = image.imread('../img/catdog.jpg')
+print('Input', img.shape)
+x = img.astype('float32').transpose((2,0,1)).expand_dims(axis=0)/255
 y = conv_trans(x)
 y = y[0].clip(0,1).transpose((1,2,0))
 print('Output', y.shape)
-
 gb.plt.imshow(y.asnumpy());
 ```
 
-所以网络的初始化包括了三部分。主体卷积网络从训练好的ResNet18复制得来，替代ResNet18最后全连接的卷积层使用随机初始化。
-
-最后的卷积转置层则使用双线性差值。对于卷积转置层，我们可以自定义一个初始化类。简单起见，这里我们直接通过权重的`set_data`函数改写权重。记得我们介绍过Gluon使用延后初始化来减少构造网络时需要制定输入大小。所以我们先随意初始化它，计算一次`forward`，然后再改写权重。
+下面对`net`的最后两层进行初始化。其中$1\times 1$卷积层使用Xavier，转置卷积层则使用双线性差值核。
 
-```{.python .input  n=18}
-conv_trans = net[-1]
-conv_trans.initialize(init=init.Zero())
+```{.python .input  n=8}
+trans_conv_weights = bilinear_kernel(num_classes, num_classes, 64)
+net[-1].initialize(init.Constant(trans_conv_weights))
 net[-2].initialize(init=init.Xavier())
-
-x = nd.zeros((batch_size, 3, *input_shape))
-net(x)
-
-shape = conv_trans.weight.data().shape
-conv_trans.weight.set_data(bilinear_kernel(*shape[0:3]))
-
 ```
 
-这时候我们可以真正开始训练了。值得一提的是我们使用卷积转置层的通道来预测像素的类别。所以在做`softmax`和预测的时候我们需要使用通道这个维度，既维度1. 所以在`SoftmaxCrossEntropyLoss`里加入了额外了`axis=1`选项。其他的部分跟之前的训练一致。
+## 训练
 
-```{.python .input  n=19}
-loss = gluon.loss.SoftmaxCrossEntropyLoss(axis=1)
+这时候我们可以真正开始训练了。我们使用较大的输入图片尺寸，其值选成了32的倍数。因为我们使用转置卷积层的通道来预测像素的类别，所以在做softmax是作用在通道这个维度（维度1），所以在`SoftmaxCrossEntropyLoss`里加入了额外了`axis=1`选项。
 
+```{.python .input  n=9}
+input_shape = (320, 480)
+batch_size = 32
 ctx = gb.try_all_gpus()
+loss = gluon.loss.SoftmaxCrossEntropyLoss(axis=1)
 net.collect_params().reset_ctx(ctx)
-
 trainer = gluon.Trainer(net.collect_params(),
                         'sgd', {'learning_rate': .1, 'wd':1e-3})
-
+train_data, test_data = gb.load_data_pascal_voc(batch_size, input_shape)
 gb.train(train_data, test_data, net, loss, trainer, ctx, num_epochs=10)
 ```
 
 ## 预测
 
-预测函数跟之前的图片分类预测类似，但跟上面一样，主要不同在于我们需要在`axis=1`上做`argmax`. 同时我们定义`image2label`的反函数，它将预测值转成图片。
+预测一张新图片时，我们只需要将其归一化并转成卷积网络需要的4D格式。
 
-```{.python .input  n=22}
+```{.python .input  n=57}
 def predict(im):
-    data = normalize_image(im)
+    data = test_data._dataset.normalize_image(im)
     data = data.transpose((2,0,1)).expand_dims(axis=0)
     yhat = net(data.as_in_context(ctx[0]))
     pred = nd.argmax(yhat, axis=1)
     return pred.reshape((pred.shape[1], pred.shape[2]))
 
+```
+
+同时我们根据每个像素预测的类别找出其RGB颜色以便画图。
+
+```{.python .input}
 def label2image(pred):
-    x = pred.astype('int32').asnumpy()
-    cm = nd.array(colormap).astype('uint8')
-    return nd.array(cm[x,:])
+    colormap = nd.array(
+        test_data._dataset.voc_colormap, ctx=ctx[0],dtype='uint8')
+    x = pred.astype('int32')
+    return colormap[x,:]
 ```
 
-我们读取前几张测试图片并对其进行预测。
+现在我们读取前几张测试图片并对其进行预测。
 
-```{.python .input  n=27}
-test_images, test_labels = read_images(train=False)
+```{.python .input  n=58}
+test_images, test_labels = gb.read_voc_images(train=False)
 
+n = 5
 imgs = []
 for i in range(n):
     x = test_images[i]
@@ -353,15 +184,20 @@ gb.show_images(imgs[::3]+imgs[1::3]+imgs[2::3], 3, n);
 
 ## 小结
 
-* 通过使用卷积转置层，我们可以得到更大分辨率的输出。
+FCN通过使用转置卷积层来为每个像素预测类别。
 
 ## 练习
 
-* 试着改改最后的卷积转置层的参数设定
-* 看看双线性差值初始化是不是必要的
-* 试着改改训练参数来使得收敛更好些
-* [FCN论文](https://arxiv.org/abs/1411.4038)中提到了不只是使用主体卷积网络输出，还可以将前面层的输出也加进来。试着实现。
+* 试着改改最后的转置卷积层的参数设定。
+* 看看双线性差值初始化是不是必要的。
+* 试着改改训练参数来使得收敛更好些。
+* FCN论文[1]中提到了不只是使用主体卷积网络输出，还可以考虑其中间层的输出。试着实现这个想法。
 
 ## 扫码直达[讨论区](https://discuss.gluon.ai/t/topic/3041)
 
 ![](../img/qr_fcn.svg)
+
+
+## 参考文献
+
+[1] Long, Jonathan, Evan Shelhamer, and Trevor Darrell. "Fully convolutional networks for semantic segmentation." CVPR. 2015.

chapter_computer-vision/index.md

@@ -9,10 +9,11 @@
    image-augmentation
    fine-tuning
    bounding-box
-   pikachu
    anchor
+   pikachu
    ssd
    rcnn
+   voc
    fcn
    neural-style
    kaggle-gluon-cifar10

chapter_computer-vision/voc.md

+# 语义分割和数据集
+
+图片分类关心识别图片里面的主要物体，物体识别则进一步找出图片的多个物体以及它们的方形边界框。本小节我们将介绍语义分割（semantic segmentation），它在物体识别上更进一步的找出物体的精确边界框。换句话说，它识别图片中的每个像素属于哪类我们感兴趣的物体还是只是背景。下图演示猫和狗图片在语义分割中的标注。可以看到，跟物体识别相比，语义分割预测的边框更加精细。
+
+![语义分割的训练数据和标注。](../img/segmentation.svg)
+
+在计算机视觉里，还有两个跟语义分割相似的任务。一个是图片分割（image segmentation），它也是将像素划分到不同的类。不同的是，语义分割里我们赋予像素语义信息，例如属于猫、狗或者背景。而图片分割则通常根据像素本身之间的相似性，它训练时不需要像素标注信息，其预测结果也不能保证有语义性。例如图片分割可能将上图中的狗划分成两个区域，其中一个嘴巴和眼睛，其颜色以黑色为主，另一个是身体其余部分，其主色调是黄色。
+
+另一个应用是实例分割（instance segementation），它不仅需要知道每个像素的语义，即属于那一类物体，还需要进一步区分物体实例。例如如果图片中有两只狗，那么对于预测为对应狗的像素是属于地一只狗还是第二只。
+
+## Pascal VOC语义分割数据集
+
+下面我们使用一个常用的语义分割数据集
+[Pascal VOC2012](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/)来介绍这个应用。
+
+```{.python .input  n=1}
+%matplotlib inline
+import sys
+sys.path.append('..')
+import gluonbook as gb
+import tarfile
+from mxnet import nd, image, gluon
+```
+
+我们首先下载这个数据集到`../data`下。压缩包大小是2GB，下载需要一定时间。解压之后这个数据集将会放置在`../data/VOCdevkit/VOC2012`下。
+
+```{.python .input  n=2}
+data_dir = '../data'
+voc_dir = data_dir + '/VOCdevkit/VOC2012'
+url = ('http://host.robots.ox.ac.uk/pascal/VOC/voc2012'
+       '/VOCtrainval_11-May-2012.tar')
+sha1 = '4e443f8a2eca6b1dac8a6c57641b67dd40621a49'
+
+fname = gluon.utils.download(url, data_dir, sha1_hash=sha1)
+
+with tarfile.open(fname, 'r') as f:
+    f.extractall(data_dir)
+```
+
+在`ImageSets/Segmentation`下有文本文件指定哪些样本用来训练，哪些用来测试。样本图片放置在`JPEGImages`下，标注则放在`SegmentationClass`下。这里标注也是图片格式，图片大小与对应的样本图片一致，其中颜色相同的像素属于同一个类。
+
+下面定义函数将图片和标注全部读进内存。
+
+```{.python .input  n=3}
+def read_voc_images(root=voc_dir, train=True):
+    txt_fname = '%s/ImageSets/Segmentation/%s'%(
+        root, 'train.txt' if train else 'val.txt')
+    with open(txt_fname, 'r') as f:
+        images = f.read().split()
+    data, label = [None] * len(images), [None] * len(images)
+    for i, fname in enumerate(images):
+        data[i] = image.imread('%s/JPEGImages/%s.jpg'%(root, fname))
+        label[i] = image.imread('%s/SegmentationClass/%s.png'%(root, fname))
+    return data, label
+
+train_images, train_labels = read_voc_images()
+```
+
+我们画出前面五张图片和它们对应的标注。在标注，白色代表边框黑色代表背景，其他不同的颜色对应不同物体类别。
+
+```{.python .input  n=4}
+n = 5
+imgs = train_images[0:n] + train_labels[0:n]
+gb.show_images(imgs, 2, n);
+```
+
+接下来我们列出标注中每个RGB颜色值对应的类别。
+
+```{.python .input  n=5}
+voc_colormap = [[0,0,0],[128,0,0],[0,128,0], [128,128,0], [0,0,128],
+                [128,0,128],[0,128,128],[128,128,128],[64,0,0],[192,0,0],
+                [64,128,0],[192,128,0],[64,0,128],[192,0,128],
+                [64,128,128],[192,128,128],[0,64,0],[128,64,0],
+                [0,192,0],[128,192,0],[0,64,128]]
+
+voc_classes = ['background','aeroplane','bicycle','bird','boat',
+               'bottle','bus','car','cat','chair','cow','diningtable',
+               'dog','horse','motorbike','person','potted plant',
+               'sheep','sofa','train','tv/monitor']
+```
+
+这样给定一个标号图片，我们就可以将每个像素对应的物体标号找出来。
+
+```{.python .input  n=6}
+colormap2label = nd.zeros(256**3)
+for i, cm in enumerate(voc_colormap):
+    colormap2label[(cm[0]*256+cm[1]) * 256 + cm[2]] = i
+
+def voc_label_indices(img):
+    data = img.astype('int32')
+    idx = (data[:,:,0]*256+data[:,:,1])*256+data[:,:,2]
+    return colormap2label[idx]
+```
+
+可以看到第一张样本中飞机头部对应的标注里属于飞机的像素被标记成了1。
+
+```{.python .input  n=7}
+y = voc_label_indices(train_labels[0])
+y[105:115, 130:140]
+```
+
+### 数据预处理
+
+我们知道小批量训练需要输入图片的形状一致。之前我们通过图片缩放来得到同样形状的输入。但语义分割里，如果对样本图片进行缩放，那么重新映射每个像素对应的类别将变得困难，特别是对应物体边缘的像素。
+
+为了避免这个困难，这里我们将图片剪裁成固定大小而不是缩放。特别的，我们使用随机剪裁来附加图片增广。下面定义随机剪裁函数，其对样本图片和标注使用用样的剪裁。
+
+```{.python .input  n=8}
+def rand_crop(data, label, height, width):
+    data, rect = image.random_crop(data, (width, height))
+    label = image.fixed_crop(label, *rect)
+    return data, label
+
+imgs = []
+for _ in range(n):
+    imgs += rand_crop(train_images[0], train_labels[0], 200, 300)
+gb.show_images(imgs[::2]+imgs[1::2], 2, n);
+```
+
+### 数据读取
+
+下面我们定义Gluon可以使用的数据集类，它可以返回任意的第$i$个样本图片和标号。除了随机剪裁外，这里我们将样本图片进行了归一化，同时过滤了小于剪裁尺寸的图片。
+
+```{.python .input  n=9}
+class VOCSegDataset(gluon.data.Dataset):
+    def __init__(self, train, crop_size):
+        self.rgb_mean = nd.array([0.485, 0.456, 0.406])
+        self.rgb_std = nd.array([0.229, 0.224, 0.225])
+        self.crop_size = crop_size        
+        data, label = read_voc_images(train=train)
+        self.data = [self.normalize_image(im) for im in self.filter(data)]
+        self.label = self.filter(label)            
+        print('Read '+str(len(self.data))+' examples')
+        
+    def normalize_image(self, data):
+        return (data.astype('float32') / 255 - self.rgb_mean) / self.rgb_std
+    
+    def filter(self, images):
+        return [im for im in images if (
+            im.shape[0] >= self.crop_size[0] and
+            im.shape[1] >= self.crop_size[1])]
+
+    def __getitem__(self, idx):
+        data, label = rand_crop(self.data[idx], self.label[idx],
+                                *self.crop_size)
+        return data.transpose((2,0,1)), voc_label_indices(label)
+
+    def __len__(self):
+        return len(self.data)
+```
+
+假设我们剪裁$320\times 480$图片来进行训练，我们可以查看训练和测试各保留了多少图片。
+
+```{.python .input  n=10}
+output_shape = (320, 480)  # 高和宽
+voc_train = VOCSegDataset(True, output_shape)
+voc_test = VOCSegDataset(False, output_shape)
+```
+
+最后定义批量读取，这里使用4个进程来加速读取（代码保存在gluonbook的`load_data_pascal_voc`函数里方便之后使用）。
+
+```{.python .input  n=11}
+batch_size = 64
+train_data = gluon.data.DataLoader(
+    voc_train, batch_size, shuffle=True,last_batch='discard', num_workers=4)
+test_data = gluon.data.DataLoader(
+    voc_test, batch_size,last_batch='discard', num_workers=4)
+```
+
+打印第一个批量可以看到，不同于图片分类和物体识别，这里的标注是一个三维的数组。
+
+```{.python .input  n=12}
+for data, label in train_data:
+    print(data.shape)
+    print(label.shape)
+    break
+```
+
+## 小节
+
+## 练习

gluonbook/utils.py

 import random
+import os
+import tarfile
 from time import time
 
 from IPython.display import set_matplotlib_formats
@@ -487,3 +489,97 @@ def load_data_pikachu(batch_size, edge_size=256):
         data_shape=(3, edge_size, edge_size),
         shuffle=False)
     return (train_data, val_data)
+
+def _download_voc_pascal(data_dir='../data'):
+    voc_dir = data_dir + '/VOCdevkit/VOC2012'
+    url = ('http://host.robots.ox.ac.uk/pascal/VOC/voc2012'
+           '/VOCtrainval_11-May-2012.tar')
+    sha1 = '4e443f8a2eca6b1dac8a6c57641b67dd40621a49'
+    fname = gluon.utils.download(url, data_dir, sha1_hash=sha1)
+    if not os.path.exists(voc_dir+'/ImageSets/Segmentation/train.txt'):
+        with tarfile.open(fname, 'r') as f:
+            f.extractall(data_dir)
+    return voc_dir
+
+def read_voc_images(root='../data/VOCdevkit/VOC2012', train=True):
+    txt_fname = '%s/ImageSets/Segmentation/%s'%(
+        root, 'train.txt' if train else 'val.txt')
+    with open(txt_fname, 'r') as f:
+        images = f.read().split()
+    data, label = [None] * len(images), [None] * len(images)
+    for i, fname in enumerate(images):
+        data[i] = image.imread('%s/JPEGImages/%s.jpg'%(root, fname))
+        label[i] = image.imread('%s/SegmentationClass/%s.png'%(root, fname))
+    return data, label
+
+
+voc_rgb_mean = nd.array([0.485, 0.456, 0.406])
+voc_rgb_std = nd.array([0.229, 0.224, 0.225])
+
+def normalize_voc_image(data):
+    return (data.astype('float32') / 255 - voc_rgb_mean) / voc_rgb_std
+
+class VOCSegDataset(gluon.data.Dataset):
+    def __init__(self, train, crop_size):
+        self.train = train
+        self.crop_size = crop_size
+        self.rgb_mean = nd.array([0.485, 0.456, 0.406])
+        self.rgb_std = nd.array([0.229, 0.224, 0.225])
+        self.voc_colormap = [[0,0,0],[128,0,0],[0,128,0], [128,128,0], [0,0,128],
+                             [128,0,128],[0,128,128],[128,128,128],[64,0,0],[192,0,0],
+                             [64,128,0],[192,128,0],[64,0,128],[192,0,128],
+                             [64,128,128],[192,128,128],[0,64,0],[128,64,0],
+                             [0,192,0],[128,192,0],[0,64,128]]
+        self.voc_classes = ['background','aeroplane','bicycle','bird','boat',
+                            'bottle','bus','car','cat','chair','cow','diningtable',
+                            'dog','horse','motorbike','person','potted plant',
+                            'sheep','sofa','train','tv/monitor']
+        self.colormap2label = None
+        self.load_images()
+
+    def voc_label_indices(self, img):
+        if self.colormap2label is None:
+            self.colormap2label = nd.zeros(256**3)
+            for i, cm in enumerate(self.voc_colormap):
+                self.colormap2label[(cm[0]*256+cm[1]) * 256 + cm[2]] = i
+        data = img.astype('int32')
+        idx = (data[:,:,0]*256+data[:,:,1])*256+data[:,:,2]
+        return self.colormap2label[idx]
+
+    def rand_crop(self, data, label, height, width):
+        data, rect = image.random_crop(data, (width, height))
+        label = image.fixed_crop(label, *rect)
+        return data, label
+
+    def load_images(self):
+        voc_dir = _download_voc_pascal()
+        data, label = read_voc_images(root=voc_dir, train=self.train)
+        self.data = [self.normalize_image(im) for im in self.filter(data)]
+        self.label = self.filter(label)
+        print('Read '+str(len(self.data))+' examples')
+
+    def normalize_image(self, data):
+        return (data.astype('float32') / 255 - self.rgb_mean) / self.rgb_std
+
+    def filter(self, images):
+        return [im for im in images if (
+            im.shape[0] >= self.crop_size[0] and
+            im.shape[1] >= self.crop_size[1])]
+
+    def __getitem__(self, idx):
+        data, label = self.rand_crop(self.data[idx], self.label[idx],
+                                *self.crop_size)
+        return data.transpose((2,0,1)), self.voc_label_indices(label)
+
+    def __len__(self):
+        return len(self.data)
+
+def load_data_pascal_voc(batch_size, output_shape):
+    voc_train = VOCSegDataset(True, output_shape)
+    voc_test = VOCSegDataset(False, output_shape)
+
+    train_data = gluon.data.DataLoader(
+        voc_train, batch_size, shuffle=True,last_batch='discard', num_workers=4)
+    test_data = gluon.data.DataLoader(
+        voc_test, batch_size,last_batch='discard', num_workers=4)
+    return train_data, test_data

img/fcn.svg

 <?xml version="1.0" encoding="UTF-8"?>
-<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="209pt" height="206pt" viewBox="0 0 209 206" version="1.1">
+<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="209pt" height="242pt" viewBox="0 0 209 242" version="1.1">
 <defs>
 <g>
 <symbol overflow="visible" id="glyph0-0">
@@ -42,31 +42,31 @@
 <path style="stroke:none;" d="M 2.921875 -3.90625 L 2.921875 -3.328125 L 7.796875 -3.328125 L 7.796875 -3.90625 Z M 2.28125 -2.4375 L 2.28125 -1.84375 L 4.25 -1.84375 C 3.625 -0.828125 3.09375 -0.203125 2.703125 0.03125 C 2.65625 0.046875 2.625 0.078125 2.5625 0.09375 L 2.703125 0.640625 C 4.40625 0.53125 5.96875 0.375 7.390625 0.1875 C 7.578125 0.40625 7.734375 0.640625 7.890625 0.875 L 8.40625 0.515625 C 8.015625 -0.0625 7.4375 -0.75 6.703125 -1.53125 L 6.21875 -1.234375 C 6.484375 -0.921875 6.75 -0.625 7 -0.328125 C 5.84375 -0.15625 4.671875 -0.03125 3.484375 0.046875 C 3.90625 -0.328125 4.390625 -0.96875 4.9375 -1.84375 L 8.390625 -1.84375 L 8.390625 -2.4375 Z M 7.296875 -6.390625 L 7.296875 -5.34375 L 2 -5.34375 L 2 -6.390625 Z M 2 -4.765625 L 7.921875 -4.765625 L 7.921875 -7 L 1.375 -7 L 1.375 -4.015625 C 1.359375 -2.203125 1.046875 -0.734375 0.421875 0.40625 L 0.890625 0.84375 C 1.625 -0.515625 2 -2.140625 2 -4.015625 Z M 2 -4.765625 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-13">
-<path style="stroke:none;" d="M 2.171875 -4.875 L 2.171875 -3.515625 L 1.1875 -3.515625 C 1.4375 -4.09375 1.671875 -4.765625 1.890625 -5.546875 L 3.765625 -5.546875 L 3.765625 -6.171875 L 2.046875 -6.171875 C 2.140625 -6.53125 2.21875 -6.890625 2.296875 -7.28125 L 1.6875 -7.375 C 1.59375 -6.953125 1.515625 -6.546875 1.421875 -6.171875 L 0.359375 -6.171875 L 0.359375 -5.546875 L 1.265625 -5.546875 C 1.03125 -4.71875 0.796875 -4.046875 0.53125 -3.5 L 0.671875 -2.921875 L 2.171875 -2.921875 L 2.171875 -1.578125 C 1.609375 -1.46875 1 -1.390625 0.34375 -1.3125 L 0.4375 -0.671875 C 1.046875 -0.75 1.625 -0.84375 2.171875 -0.9375 L 2.171875 0.890625 L 2.796875 0.890625 L 2.796875 -1.0625 C 3.234375 -1.15625 3.640625 -1.265625 4.046875 -1.375 L 4.046875 -1.984375 C 3.65625 -1.875 3.234375 -1.78125 2.796875 -1.6875 L 2.796875 -2.921875 L 3.796875 -2.921875 L 3.796875 -3.515625 L 2.796875 -3.515625 L 2.796875 -4.875 Z M 4.171875 -6.171875 L 4.171875 -5.546875 L 5.46875 -5.546875 C 5.390625 -5.203125 5.3125 -4.859375 5.234375 -4.515625 L 3.921875 -4.515625 L 3.921875 -3.890625 L 5.078125 -3.890625 C 4.921875 -3.296875 4.75 -2.734375 4.578125 -2.171875 L 7.59375 -2.171875 C 7.265625 -1.640625 6.84375 -1.15625 6.3125 -0.671875 C 5.890625 -0.90625 5.453125 -1.140625 4.984375 -1.328125 L 4.640625 -0.8125 C 5.640625 -0.40625 6.59375 0.171875 7.53125 0.890625 L 7.890625 0.34375 C 7.5625 0.09375 7.203125 -0.140625 6.84375 -0.375 C 7.484375 -0.984375 7.953125 -1.59375 8.25 -2.203125 L 8.25 -2.75 L 5.375 -2.75 C 5.5 -3.09375 5.609375 -3.46875 5.71875 -3.890625 L 8.65625 -3.890625 L 8.65625 -4.515625 L 5.875 -4.515625 L 6.109375 -5.546875 L 8.421875 -5.546875 L 8.421875 -6.171875 L 6.234375 -6.171875 C 6.3125 -6.546875 6.375 -6.9375 6.4375 -7.328125 L 5.8125 -7.390625 C 5.734375 -6.96875 5.65625 -6.5625 5.59375 -6.171875 Z M 4.171875 -6.171875 "/>
+<path style="stroke:none;" d="M 2.0625 -6.421875 C 1.859375 -6.1875 1.609375 -5.96875 1.296875 -5.765625 C 0.984375 -5.59375 0.6875 -5.46875 0.390625 -5.375 L 0.390625 -4.640625 C 1 -4.828125 1.5 -5.109375 1.90625 -5.53125 L 1.90625 0 L 2.640625 0 L 2.640625 -6.421875 Z M 2.0625 -6.421875 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-14">
-<path style="stroke:none;" d="M 0.640625 -4.8125 L 0.640625 -4.3125 L 4.0625 -4.3125 C 4.03125 -4.140625 4 -3.96875 3.96875 -3.796875 L 1.78125 -3.796875 L 1.78125 0.171875 L 0.484375 0.171875 L 0.484375 0.6875 L 8.53125 0.6875 L 8.53125 0.171875 L 7.234375 0.171875 L 7.234375 -3.796875 L 4.609375 -3.796875 C 4.65625 -3.953125 4.6875 -4.125 4.71875 -4.3125 L 8.375 -4.3125 L 8.375 -4.8125 L 4.78125 -4.8125 L 4.84375 -5.328125 L 7.765625 -5.328125 L 7.765625 -7.015625 L 1.234375 -7.015625 L 1.234375 -5.328125 L 4.1875 -5.328125 L 4.140625 -4.8125 Z M 2.40625 0.171875 L 2.40625 -0.421875 L 6.59375 -0.421875 L 6.59375 0.171875 Z M 2.40625 -0.828125 L 2.40625 -1.390625 L 6.59375 -1.390625 L 6.59375 -0.828125 Z M 2.40625 -1.8125 L 2.40625 -2.375 L 6.59375 -2.375 L 6.59375 -1.8125 Z M 2.40625 -2.78125 L 2.40625 -3.34375 L 6.59375 -3.34375 L 6.59375 -2.78125 Z M 7.140625 -5.78125 L 5.8125 -5.78125 L 5.8125 -6.53125 L 7.140625 -6.53125 Z M 5.21875 -5.78125 L 3.78125 -5.78125 L 3.78125 -6.53125 L 5.21875 -6.53125 Z M 3.203125 -5.78125 L 1.859375 -5.78125 L 1.859375 -6.53125 L 3.203125 -6.53125 Z M 3.203125 -5.78125 "/>
+<path style="stroke:none;" d=""/>
 </symbol>
 <symbol overflow="visible" id="glyph0-15">
-<path style="stroke:none;" d="M 0.515625 -7.4375 C 0.875 -6.84375 1.171875 -6.234375 1.375 -5.578125 C 1.640625 -4.78125 1.765625 -4.015625 1.765625 -3.234375 C 1.765625 -2.46875 1.640625 -1.6875 1.375 -0.90625 C 1.171875 -0.265625 0.875 0.359375 0.515625 0.96875 L 1.15625 0.96875 C 1.53125 0.375 1.84375 -0.25 2.078125 -0.890625 C 2.34375 -1.671875 2.46875 -2.453125 2.46875 -3.234375 C 2.46875 -4.03125 2.34375 -4.8125 2.078125 -5.59375 C 1.84375 -6.25 1.53125 -6.859375 1.15625 -7.4375 Z M 0.515625 -7.4375 "/>
+<path style="stroke:none;" d="M 0.25 -4.65625 L 1.890625 -2.46875 L 0.046875 0 L 0.90625 0 L 2.296875 -1.9375 L 3.6875 0 L 4.53125 0 L 2.6875 -2.46875 L 4.34375 -4.65625 L 3.484375 -4.65625 L 2.296875 -3 L 1.09375 -4.65625 Z M 0.25 -4.65625 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-16">
-<path style="stroke:none;" d="M 0.453125 -3.5625 L 0.453125 -2.953125 L 5 -2.953125 L 5 -3.5625 Z M 0.453125 -1.703125 L 0.453125 -1.09375 L 5 -1.09375 L 5 -1.703125 Z M 0.453125 -1.703125 "/>
+<path style="stroke:none;" d="M 0.515625 -7.4375 C 0.875 -6.84375 1.171875 -6.234375 1.375 -5.578125 C 1.640625 -4.78125 1.765625 -4.015625 1.765625 -3.234375 C 1.765625 -2.46875 1.640625 -1.6875 1.375 -0.90625 C 1.171875 -0.265625 0.875 0.359375 0.515625 0.96875 L 1.15625 0.96875 C 1.53125 0.375 1.84375 -0.25 2.078125 -0.890625 C 2.34375 -1.671875 2.46875 -2.453125 2.46875 -3.234375 C 2.46875 -4.03125 2.34375 -4.8125 2.078125 -5.59375 C 1.84375 -6.25 1.53125 -6.859375 1.15625 -7.4375 Z M 0.515625 -7.4375 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-17">
-<path style="stroke:none;" d="M 2.703125 -6.546875 C 1.921875 -6.546875 1.34375 -6.21875 0.96875 -5.53125 C 0.625 -4.953125 0.453125 -4.1875 0.453125 -3.21875 C 0.453125 -2.234375 0.625 -1.46875 0.96875 -0.890625 C 1.34375 -0.21875 1.921875 0.125 2.703125 0.125 C 3.46875 0.125 4.046875 -0.21875 4.4375 -0.890625 C 4.78125 -1.46875 4.953125 -2.234375 4.953125 -3.21875 C 4.953125 -4.1875 4.78125 -4.953125 4.4375 -5.53125 C 4.046875 -6.21875 3.46875 -6.546875 2.703125 -6.546875 Z M 2.703125 -5.921875 C 3.28125 -5.921875 3.6875 -5.609375 3.9375 -5 C 4.125 -4.5625 4.21875 -3.96875 4.21875 -3.21875 C 4.21875 -2.46875 4.125 -1.875 3.9375 -1.4375 C 3.6875 -0.8125 3.28125 -0.5 2.703125 -0.5 C 2.125 -0.5 1.703125 -0.8125 1.453125 -1.4375 C 1.28125 -1.875 1.1875 -2.46875 1.1875 -3.21875 C 1.1875 -3.96875 1.28125 -4.5625 1.453125 -5 C 1.703125 -5.609375 2.125 -5.921875 2.703125 -5.921875 Z M 2.703125 -5.921875 "/>
+<path style="stroke:none;" d="M 0.453125 -3.5625 L 0.453125 -2.953125 L 5 -2.953125 L 5 -3.5625 Z M 0.453125 -1.703125 L 0.453125 -1.09375 L 5 -1.09375 L 5 -1.703125 Z M 0.453125 -1.703125 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-18">
-<path style="stroke:none;" d="M 1.1875 -1.078125 C 1.03125 -1.078125 0.90625 -1.015625 0.796875 -0.90625 C 0.6875 -0.796875 0.640625 -0.671875 0.640625 -0.515625 C 0.640625 -0.359375 0.6875 -0.234375 0.796875 -0.125 C 0.90625 -0.015625 1.03125 0.03125 1.1875 0.03125 C 1.328125 0.03125 1.46875 -0.015625 1.578125 -0.125 C 1.6875 -0.234375 1.734375 -0.359375 1.734375 -0.515625 C 1.734375 -0.671875 1.6875 -0.796875 1.578125 -0.90625 C 1.46875 -1.015625 1.328125 -1.078125 1.1875 -1.078125 Z M 1.1875 -1.078125 "/>
+<path style="stroke:none;" d="M 2.703125 -6.546875 C 1.921875 -6.546875 1.34375 -6.21875 0.96875 -5.53125 C 0.625 -4.953125 0.453125 -4.1875 0.453125 -3.21875 C 0.453125 -2.234375 0.625 -1.46875 0.96875 -0.890625 C 1.34375 -0.21875 1.921875 0.125 2.703125 0.125 C 3.46875 0.125 4.046875 -0.21875 4.4375 -0.890625 C 4.78125 -1.46875 4.953125 -2.234375 4.953125 -3.21875 C 4.953125 -4.1875 4.78125 -4.953125 4.4375 -5.53125 C 4.046875 -6.21875 3.46875 -6.546875 2.703125 -6.546875 Z M 2.703125 -5.921875 C 3.28125 -5.921875 3.6875 -5.609375 3.9375 -5 C 4.125 -4.5625 4.21875 -3.96875 4.21875 -3.21875 C 4.21875 -2.46875 4.125 -1.875 3.9375 -1.4375 C 3.6875 -0.8125 3.28125 -0.5 2.703125 -0.5 C 2.125 -0.5 1.703125 -0.8125 1.453125 -1.4375 C 1.28125 -1.875 1.1875 -2.46875 1.1875 -3.21875 C 1.1875 -3.96875 1.28125 -4.5625 1.453125 -5 C 1.703125 -5.609375 2.125 -5.921875 2.703125 -5.921875 Z M 2.703125 -5.921875 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-19">
-<path style="stroke:none;" d="M 3.453125 -6.421875 L 0.21875 -2.15625 L 0.21875 -1.453125 L 3.453125 -1.453125 L 3.453125 0 L 4.15625 0 L 4.15625 -1.453125 L 5.1875 -1.453125 L 5.1875 -2.0625 L 4.15625 -2.0625 L 4.15625 -6.421875 Z M 3.421875 -5.46875 L 3.453125 -5.46875 L 3.453125 -2.0625 L 0.84375 -2.0625 Z M 3.421875 -5.46875 "/>
+<path style="stroke:none;" d="M 1.1875 -1.078125 C 1.03125 -1.078125 0.90625 -1.015625 0.796875 -0.90625 C 0.6875 -0.796875 0.640625 -0.671875 0.640625 -0.515625 C 0.640625 -0.359375 0.6875 -0.234375 0.796875 -0.125 C 0.90625 -0.015625 1.03125 0.03125 1.1875 0.03125 C 1.328125 0.03125 1.46875 -0.015625 1.578125 -0.125 C 1.6875 -0.234375 1.734375 -0.359375 1.734375 -0.515625 C 1.734375 -0.671875 1.6875 -0.796875 1.578125 -0.90625 C 1.46875 -1.015625 1.328125 -1.078125 1.1875 -1.078125 Z M 1.1875 -1.078125 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-20">
-<path style="stroke:none;" d="M 1.1875 -1.078125 C 1.015625 -1.078125 0.875 -1.03125 0.78125 -0.921875 C 0.65625 -0.8125 0.609375 -0.6875 0.609375 -0.515625 C 0.609375 -0.359375 0.65625 -0.21875 0.765625 -0.109375 C 0.875 -0.015625 1 0.046875 1.15625 0.046875 C 1.234375 0.046875 1.3125 0.03125 1.34375 0.015625 C 1.34375 0.21875 1.28125 0.421875 1.15625 0.609375 C 1.015625 0.796875 0.8125 0.921875 0.578125 1.015625 L 0.578125 1.453125 C 0.96875 1.34375 1.28125 1.15625 1.515625 0.859375 C 1.75 0.5625 1.859375 0.203125 1.859375 -0.1875 C 1.859375 -0.453125 1.796875 -0.65625 1.6875 -0.8125 C 1.5625 -0.984375 1.390625 -1.078125 1.1875 -1.078125 Z M 1.1875 -1.078125 "/>
+<path style="stroke:none;" d="M 3.453125 -6.421875 L 0.21875 -2.15625 L 0.21875 -1.453125 L 3.453125 -1.453125 L 3.453125 0 L 4.15625 0 L 4.15625 -1.453125 L 5.1875 -1.453125 L 5.1875 -2.0625 L 4.15625 -2.0625 L 4.15625 -6.421875 Z M 3.421875 -5.46875 L 3.453125 -5.46875 L 3.453125 -2.0625 L 0.84375 -2.0625 Z M 3.421875 -5.46875 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-21">
-<path style="stroke:none;" d=""/>
+<path style="stroke:none;" d="M 1.1875 -1.078125 C 1.015625 -1.078125 0.875 -1.03125 0.78125 -0.921875 C 0.65625 -0.8125 0.609375 -0.6875 0.609375 -0.515625 C 0.609375 -0.359375 0.65625 -0.21875 0.765625 -0.109375 C 0.875 -0.015625 1 0.046875 1.15625 0.046875 C 1.234375 0.046875 1.3125 0.03125 1.34375 0.015625 C 1.34375 0.21875 1.28125 0.421875 1.15625 0.609375 C 1.015625 0.796875 0.8125 0.921875 0.578125 1.015625 L 0.578125 1.453125 C 0.96875 1.34375 1.28125 1.15625 1.515625 0.859375 C 1.75 0.5625 1.859375 0.203125 1.859375 -0.1875 C 1.859375 -0.453125 1.796875 -0.65625 1.6875 -0.8125 C 1.5625 -0.984375 1.390625 -1.078125 1.1875 -1.078125 Z M 1.1875 -1.078125 "/>
 </symbol>
 <symbol overflow="visible" id="glyph0-22">
 <path style="stroke:none;" d="M 2.78125 -4.78125 C 2.078125 -4.78125 1.5625 -4.484375 1.234375 -3.890625 L 1.234375 -4.65625 L 0.5625 -4.65625 L 0.5625 1.78125 L 1.28125 1.78125 L 1.28125 -0.84375 C 1.640625 -0.203125 2.15625 0.125 2.8125 0.125 C 3.484375 0.125 4 -0.125 4.375 -0.609375 C 4.71875 -1.0625 4.890625 -1.625 4.890625 -2.3125 C 4.890625 -3.015625 4.71875 -3.578125 4.359375 -4.03125 C 3.984375 -4.53125 3.453125 -4.78125 2.78125 -4.78125 Z M 2.6875 -4.1875 C 3.1875 -4.1875 3.5625 -4 3.828125 -3.625 C 4.046875 -3.296875 4.15625 -2.859375 4.15625 -2.3125 C 4.15625 -1.75 4.046875 -1.3125 3.8125 -0.984375 C 3.546875 -0.640625 3.171875 -0.46875 2.65625 -0.46875 C 2.234375 -0.46875 1.90625 -0.640625 1.640625 -0.96875 C 1.375 -1.296875 1.25 -1.734375 1.25 -2.265625 L 1.25 -2.359375 C 1.25 -2.875 1.375 -3.3125 1.609375 -3.640625 C 1.859375 -4 2.21875 -4.1875 2.6875 -4.1875 Z M 2.6875 -4.1875 "/>
@@ -77,6 +77,12 @@
 <symbol overflow="visible" id="glyph0-24">
 <path style="stroke:none;" d="M 1.5 -1.078125 C 1.34375 -1.078125 1.21875 -1.03125 1.109375 -0.921875 C 1.015625 -0.8125 0.953125 -0.6875 0.953125 -0.546875 C 0.953125 -0.390625 1.015625 -0.265625 1.109375 -0.15625 C 1.21875 -0.046875 1.34375 0 1.5 0 C 1.640625 0 1.78125 -0.046875 1.890625 -0.15625 C 1.984375 -0.25 2.046875 -0.375 2.046875 -0.546875 C 2.046875 -0.6875 1.984375 -0.8125 1.890625 -0.921875 C 1.78125 -1.03125 1.65625 -1.078125 1.5 -1.078125 Z M 4.5 -1.078125 C 4.34375 -1.078125 4.21875 -1.03125 4.109375 -0.921875 C 4 -0.8125 3.953125 -0.6875 3.953125 -0.546875 C 3.953125 -0.390625 4 -0.265625 4.109375 -0.15625 C 4.21875 -0.046875 4.34375 0 4.5 0 C 4.65625 0 4.78125 -0.046875 4.890625 -0.15625 C 5 -0.25 5.046875 -0.375 5.046875 -0.546875 C 5.046875 -0.6875 5 -0.8125 4.890625 -0.921875 C 4.78125 -1.03125 4.65625 -1.078125 4.5 -1.078125 Z M 7.5 -1.078125 C 7.34375 -1.078125 7.21875 -1.03125 7.109375 -0.921875 C 7 -0.8125 6.953125 -0.6875 6.953125 -0.546875 C 6.953125 -0.390625 7 -0.265625 7.109375 -0.15625 C 7.21875 -0.046875 7.34375 0 7.5 0 C 7.65625 0 7.78125 -0.046875 7.890625 -0.15625 C 7.984375 -0.25 8.046875 -0.375 8.046875 -0.546875 C 8.046875 -0.6875 7.984375 -0.8125 7.890625 -0.921875 C 7.78125 -1.03125 7.65625 -1.078125 7.5 -1.078125 Z M 7.5 -1.078125 "/>
 </symbol>
+<symbol overflow="visible" id="glyph0-25">
+<path style="stroke:none;" d="M 2.171875 -4.875 L 2.171875 -3.515625 L 1.1875 -3.515625 C 1.4375 -4.09375 1.671875 -4.765625 1.890625 -5.546875 L 3.765625 -5.546875 L 3.765625 -6.171875 L 2.046875 -6.171875 C 2.140625 -6.53125 2.21875 -6.890625 2.296875 -7.28125 L 1.6875 -7.375 C 1.59375 -6.953125 1.515625 -6.546875 1.421875 -6.171875 L 0.359375 -6.171875 L 0.359375 -5.546875 L 1.265625 -5.546875 C 1.03125 -4.71875 0.796875 -4.046875 0.53125 -3.5 L 0.671875 -2.921875 L 2.171875 -2.921875 L 2.171875 -1.578125 C 1.609375 -1.46875 1 -1.390625 0.34375 -1.3125 L 0.4375 -0.671875 C 1.046875 -0.75 1.625 -0.84375 2.171875 -0.9375 L 2.171875 0.890625 L 2.796875 0.890625 L 2.796875 -1.0625 C 3.234375 -1.15625 3.640625 -1.265625 4.046875 -1.375 L 4.046875 -1.984375 C 3.65625 -1.875 3.234375 -1.78125 2.796875 -1.6875 L 2.796875 -2.921875 L 3.796875 -2.921875 L 3.796875 -3.515625 L 2.796875 -3.515625 L 2.796875 -4.875 Z M 4.171875 -6.171875 L 4.171875 -5.546875 L 5.46875 -5.546875 C 5.390625 -5.203125 5.3125 -4.859375 5.234375 -4.515625 L 3.921875 -4.515625 L 3.921875 -3.890625 L 5.078125 -3.890625 C 4.921875 -3.296875 4.75 -2.734375 4.578125 -2.171875 L 7.59375 -2.171875 C 7.265625 -1.640625 6.84375 -1.15625 6.3125 -0.671875 C 5.890625 -0.90625 5.453125 -1.140625 4.984375 -1.328125 L 4.640625 -0.8125 C 5.640625 -0.40625 6.59375 0.171875 7.53125 0.890625 L 7.890625 0.34375 C 7.5625 0.09375 7.203125 -0.140625 6.84375 -0.375 C 7.484375 -0.984375 7.953125 -1.59375 8.25 -2.203125 L 8.25 -2.75 L 5.375 -2.75 C 5.5 -3.09375 5.609375 -3.46875 5.71875 -3.890625 L 8.65625 -3.890625 L 8.65625 -4.515625 L 5.875 -4.515625 L 6.109375 -5.546875 L 8.421875 -5.546875 L 8.421875 -6.171875 L 6.234375 -6.171875 C 6.3125 -6.546875 6.375 -6.9375 6.4375 -7.328125 L 5.8125 -7.390625 C 5.734375 -6.96875 5.65625 -6.5625 5.59375 -6.171875 Z M 4.171875 -6.171875 "/>
+</symbol>
+<symbol overflow="visible" id="glyph0-26">
+<path style="stroke:none;" d="M 0.640625 -4.8125 L 0.640625 -4.3125 L 4.0625 -4.3125 C 4.03125 -4.140625 4 -3.96875 3.96875 -3.796875 L 1.78125 -3.796875 L 1.78125 0.171875 L 0.484375 0.171875 L 0.484375 0.6875 L 8.53125 0.6875 L 8.53125 0.171875 L 7.234375 0.171875 L 7.234375 -3.796875 L 4.609375 -3.796875 C 4.65625 -3.953125 4.6875 -4.125 4.71875 -4.3125 L 8.375 -4.3125 L 8.375 -4.8125 L 4.78125 -4.8125 L 4.84375 -5.328125 L 7.765625 -5.328125 L 7.765625 -7.015625 L 1.234375 -7.015625 L 1.234375 -5.328125 L 4.1875 -5.328125 L 4.140625 -4.8125 Z M 2.40625 0.171875 L 2.40625 -0.421875 L 6.59375 -0.421875 L 6.59375 0.171875 Z M 2.40625 -0.828125 L 2.40625 -1.390625 L 6.59375 -1.390625 L 6.59375 -0.828125 Z M 2.40625 -1.8125 L 2.40625 -2.375 L 6.59375 -2.375 L 6.59375 -1.8125 Z M 2.40625 -2.78125 L 2.40625 -3.34375 L 6.59375 -3.34375 L 6.59375 -2.78125 Z M 7.140625 -5.78125 L 5.8125 -5.78125 L 5.8125 -6.53125 L 7.140625 -6.53125 Z M 5.21875 -5.78125 L 3.78125 -5.78125 L 3.78125 -6.53125 L 5.21875 -6.53125 Z M 3.203125 -5.78125 L 1.859375 -5.78125 L 1.859375 -6.53125 L 3.203125 -6.53125 Z M 3.203125 -5.78125 "/>
+</symbol>
 <symbol overflow="visible" id="glyph1-0">
 <path style="stroke:none;" d="M 3.6875 -6 L 0.921875 -6 L 0.921875 -0.421875 L 3.6875 -0.421875 Z M 4.140625 -6.40625 L 4.140625 -0.015625 L 0.453125 -0.015625 L 0.453125 -6.40625 Z M 4.140625 -6.40625 "/>
 </symbol>
@@ -88,22 +94,22 @@
 </symbol>
 </g>
 <clipPath id="clip1">
-  <path d="M 136 148 L 208 148 L 208 205.101562 L 136 205.101562 Z M 136 148 "/>
+  <path d="M 136 184 L 208 184 L 208 241.101562 L 136 241.101562 Z M 136 184 "/>
 </clipPath>
 <image id="image5" width="728" height="561" xlink:href="data:image/jpeg;base64,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"/>
 <clipPath id="clip2">
-  <path d="M 1 148 L 73 148 L 73 205.101562 L 1 205.101562 Z M 1 148 "/>
+  <path d="M 1 184 L 73 184 L 73 241.101562 L 1 241.101562 Z M 1 184 "/>
 </clipPath>
 <image id="image8" width="728" height="561" xlink:href="data:image/jpeg;base64,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"/>
 </defs>
 <g id="surface1">
-<rect x="0" y="0" width="209" height="206" style="fill:rgb(100%,100%,100%);fill-opacity:1;stroke:none;"/>
-<rect x="0" y="0" width="209" height="206" style="fill:rgb(100%,100%,100%);fill-opacity:1;stroke:none;"/>
+<rect x="0" y="0" width="209" height="242" style="fill:rgb(100%,100%,100%);fill-opacity:1;stroke:none;"/>
+<rect x="0" y="0" width="209" height="242" style="fill:rgb(100%,100%,100%);fill-opacity:1;stroke:none;"/>
 <g clip-path="url(#clip1)" clip-rule="nonzero">
-<use xlink:href="#image5" transform="matrix(0.0989011,0,0,0.101786,136,148.000002)"/>
+<use xlink:href="#image5" transform="matrix(0.0989011,0,0,0.101786,136,184.000002)"/>
 </g>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 252.5 L 270.5 252.5 L 270.5 309.5 L 198.5 309.5 Z M 198.5 252.5 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 105.5 L 270.5 105.5 L 270.5 162.5 L 198.5 162.5 Z M 198.5 105.5 " transform="matrix(1,0,0,1,-62,-104)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 252.5 L 270.5 252.5 L 270.5 309.5 L 198.5 309.5 Z M 198.5 252.5 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 69.5 L 270.5 69.5 L 270.5 126.5 L 198.5 126.5 Z M 198.5 69.5 " transform="matrix(1,0,0,1,-62,-68)"/>
 <path style=" stroke:none;fill-rule:nonzero;fill:rgb(82.469177%,90.846252%,100%);fill-opacity:1;" d="M 149.960938 6.179688 L 148.378906 6.6875 C 148.328125 6.703125 148.273438 6.726562 148.226562 6.75 L 146.394531 7.695312 C 146.351562 7.71875 146.3125 7.742188 146.269531 7.773438 L 144.96875 8.707031 C 144.820312 8.816406 144.703125 8.960938 144.632812 9.128906 L 144.0625 10.480469 C 144.050781 10.507812 144.035156 10.539062 144.023438 10.566406 L 143.605469 11.335938 C 143.601562 11.34375 143.597656 11.355469 143.59375 11.363281 L 143.105469 12.347656 C 143.035156 12.484375 143 12.636719 143 12.792969 L 143 14.582031 C 143 14.628906 142.996094 14.675781 142.992188 14.722656 L 142.652344 17.140625 C 142.648438 17.175781 142.644531 17.210938 142.644531 17.246094 L 142.589844 19 C 142.585938 19.105469 142.597656 19.214844 142.628906 19.320312 L 142.910156 20.25 C 142.945312 20.371094 142.957031 20.496094 142.949219 20.621094 L 142.921875 20.933594 C 142.910156 21.105469 142.9375 21.277344 143.011719 21.433594 L 143.242188 21.933594 C 143.269531 21.992188 143.300781 22.046875 143.335938 22.097656 L 143.894531 22.882812 C 143.910156 22.898438 143.921875 22.921875 143.9375 22.941406 L 144.59375 24.015625 C 144.609375 24.042969 144.625 24.066406 144.640625 24.085938 L 145.152344 24.785156 C 145.257812 24.929688 145.324219 25.101562 145.339844 25.277344 L 145.441406 26.273438 C 145.441406 26.273438 145.441406 26.277344 145.441406 26.28125 L 145.617188 28.191406 C 145.625 28.3125 145.613281 28.433594 145.582031 28.550781 L 145.203125 29.902344 C 145.191406 29.941406 145.179688 29.980469 145.164062 30.015625 L 144.339844 32.019531 C 144.3125 32.085938 144.292969 32.15625 144.28125 32.230469 L 144.003906 33.832031 C 143.988281 33.917969 143.984375 34.007812 143.992188 34.09375 L 144.230469 36.722656 C 144.234375 36.746094 144.238281 36.769531 144.238281 36.792969 L 144.496094 38.378906 C 144.53125 38.601562 144.492188 38.835938 144.378906 39.03125 L 143.316406 40.886719 C 143.292969 40.929688 143.265625 40.96875 143.238281 41.003906 L 141.949219 42.660156 C 141.949219 42.664062 141.945312 42.664062 141.945312 42.667969 L 141.316406 43.492188 C 141.191406 43.65625 141.117188 43.855469 141.109375 44.0625 L 141.058594 45.414062 C 141.058594 45.445312 141.058594 45.472656 141.050781 45.503906 L 140.898438 46.695312 C 140.878906 46.84375 140.824219 46.984375 140.742188 47.109375 L 140.160156 48.003906 C 140.054688 48.167969 140 48.355469 140 48.550781 L 140 48.792969 C 140 49.113281 140.214844 49.390625 140.519531 49.476562 L 141.160156 49.65625 C 141.300781 49.695312 141.433594 49.765625 141.546875 49.863281 L 142.167969 50.402344 C 142.273438 50.492188 142.398438 50.5625 142.535156 50.601562 L 143.710938 50.957031 C 143.804688 50.984375 143.902344 51 144 51 L 145.761719 51 C 145.800781 51 145.84375 51.003906 145.882812 51.007812 L 147.476562 51.207031 C 147.640625 51.226562 147.796875 51.289062 147.933594 51.386719 L 149.027344 52.167969 C 149.050781 52.183594 149.074219 52.199219 149.097656 52.214844 L 150.359375 52.960938 C 150.425781 53 150.496094 53.027344 150.570312 53.050781 L 151.644531 53.390625 C 151.660156 53.398438 151.671875 53.402344 151.6875 53.40625 L 153.152344 53.941406 C 153.261719 53.980469 153.378906 54 153.492188 54 L 155.390625 54 C 155.570312 54 155.75 53.949219 155.90625 53.855469 L 156.738281 53.355469 C 156.828125 53.300781 156.925781 53.257812 157.03125 53.234375 L 157.289062 53.175781 C 157.546875 53.117188 157.765625 52.9375 157.878906 52.695312 L 157.976562 52.488281 C 158.042969 52.335938 158.074219 52.167969 158.066406 52 L 158.023438 51.410156 C 158.019531 51.339844 158.023438 51.265625 158.035156 51.195312 L 158.238281 49.8125 C 158.261719 49.664062 158.25 49.511719 158.207031 49.371094 L 158.144531 49.175781 C 158.089844 48.996094 158.191406 48.804688 158.371094 48.75 C 158.394531 48.742188 158.421875 48.738281 158.449219 48.738281 L 160.148438 48.636719 C 160.167969 48.636719 160.1875 48.632812 160.207031 48.632812 L 162.335938 48.382812 C 162.335938 48.378906 162.335938 48.378906 162.339844 48.378906 L 164.398438 48.128906 C 164.421875 48.128906 164.445312 48.125 164.46875 48.125 L 167.949219 47.9375 C 167.976562 47.9375 168.003906 47.9375 168.03125 47.9375 L 170.082031 47.988281 C 170.367188 47.996094 170.636719 48.125 170.820312 48.34375 L 171.144531 48.722656 C 171.199219 48.792969 171.265625 48.851562 171.339844 48.902344 L 172.003906 49.359375 C 172.039062 49.386719 172.074219 49.40625 172.113281 49.425781 L 173.625 50.210938 C 173.636719 50.21875 173.648438 50.226562 173.664062 50.230469 L 175.191406 50.941406 C 175.34375 51.011719 175.507812 51.042969 175.675781 51.03125 L 177.511719 50.921875 C 177.515625 50.921875 177.519531 50.921875 177.523438 50.921875 L 179.898438 50.800781 C 179.960938 50.796875 180.027344 50.785156 180.089844 50.769531 L 180.863281 50.578125 C 181.257812 50.476562 181.554688 50.148438 181.609375 49.742188 L 181.679688 49.246094 C 181.699219 49.09375 181.683594 48.941406 181.636719 48.796875 L 181.339844 47.902344 C 181.25 47.640625 181.277344 47.355469 181.40625 47.109375 L 181.425781 47.078125 C 181.566406 46.816406 181.578125 46.507812 181.460938 46.234375 L 181.339844 45.953125 C 181.292969 45.835938 181.21875 45.730469 181.128906 45.640625 L 180.339844 44.851562 C 180.238281 44.75 180.117188 44.671875 179.984375 44.621094 L 178.777344 44.167969 C 178.699219 44.140625 178.621094 44.121094 178.539062 44.109375 L 177.660156 44.007812 C 177.273438 43.964844 176.953125 43.675781 176.871094 43.292969 L 176.78125 42.894531 C 176.746094 42.734375 176.75 42.570312 176.792969 42.417969 L 176.964844 41.785156 C 176.988281 41.699219 177 41.609375 177 41.523438 L 177 40.929688 C 177 40.816406 176.980469 40.699219 176.941406 40.59375 L 176.589844 39.605469 C 176.488281 39.324219 176.265625 39.101562 175.980469 39.007812 L 175.707031 38.917969 C 175.589844 38.878906 175.484375 38.820312 175.386719 38.742188 L 174.582031 38.082031 C 174.484375 38.003906 174.402344 37.902344 174.339844 37.789062 L 173.851562 36.894531 C 173.839844 36.871094 173.828125 36.851562 173.816406 36.828125 L 173.082031 35.207031 C 173.035156 35.097656 173.003906 34.980469 172.996094 34.863281 L 172.824219 32.484375 C 172.820312 32.410156 172.808594 32.339844 172.785156 32.269531 L 172.105469 29.992188 C 172.070312 29.886719 172.023438 29.789062 171.957031 29.699219 L 170.664062 27.898438 C 170.617188 27.832031 170.5625 27.773438 170.503906 27.722656 L 168.855469 26.308594 C 168.828125 26.289062 168.800781 26.265625 168.773438 26.246094 L 167.597656 25.433594 C 167.460938 25.335938 167.425781 25.148438 167.519531 25.011719 C 167.550781 24.96875 167.59375 24.933594 167.640625 24.910156 L 167.832031 24.820312 C 168.019531 24.734375 168.171875 24.59375 168.273438 24.417969 L 168.867188 23.398438 C 168.953125 23.246094 169 23.070312 169 22.894531 L 169 21.515625 C 169 21.386719 168.976562 21.261719 168.929688 21.144531 L 167.875 18.492188 C 167.867188 18.472656 167.859375 18.453125 167.851562 18.4375 L 166.867188 16.347656 C 166.855469 16.320312 166.839844 16.289062 166.824219 16.261719 L 165.726562 14.414062 C 165.667969 14.320312 165.597656 14.234375 165.511719 14.164062 L 163.929688 12.8125 C 163.925781 12.8125 163.921875 12.808594 163.921875 12.808594 L 162.9375 11.988281 C 162.835938 11.902344 162.714844 11.839844 162.589844 11.800781 L 161.707031 11.53125 C 161.457031 11.457031 161.25 11.289062 161.125 11.0625 L 160.507812 9.964844 C 160.457031 9.871094 160.390625 9.789062 160.3125 9.71875 L 158.917969 8.433594 C 158.867188 8.386719 158.816406 8.347656 158.757812 8.3125 L 156.816406 7.136719 C 156.769531 7.109375 156.722656 7.082031 156.671875 7.0625 L 154.660156 6.257812 C 154.570312 6.222656 154.476562 6.199219 154.378906 6.191406 L 152.332031 6.003906 C 152.3125 6 152.289062 6 152.269531 6 C 152.25 6 152.226562 6 152.207031 6.003906 L 150.203125 6.132812 C 150.121094 6.136719 150.039062 6.152344 149.960938 6.179688 Z M 149.960938 6.179688 "/>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
   <use xlink:href="#glyph0-1" x="156.4978" y="33.2065"/>
@@ -118,79 +124,93 @@
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
   <use xlink:href="#glyph0-4" x="171.843651" y="24.2949"/>
 </g>
-<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 216.5 L 270.5 216.5 L 270.5 234.5 L 198.5 234.5 Z M 198.5 216.5 " transform="matrix(1,0,0,1,-62,-104)"/>
+<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 216.5 L 270.5 216.5 L 270.5 234.5 L 198.5 234.5 Z M 198.5 216.5 " transform="matrix(1,0,0,1,-62,-68)"/>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-5" x="154" y="124.5"/>
-  <use xlink:href="#glyph0-6" x="163" y="124.5"/>
-  <use xlink:href="#glyph0-7" x="172" y="124.5"/>
-  <use xlink:href="#glyph0-8" x="181" y="124.5"/>
+  <use xlink:href="#glyph0-5" x="154" y="160.5"/>
+  <use xlink:href="#glyph0-6" x="163" y="160.5"/>
+  <use xlink:href="#glyph0-7" x="172" y="160.5"/>
+  <use xlink:href="#glyph0-8" x="181" y="160.5"/>
 </g>
-<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 216.5 L 135.5 216.5 L 135.5 234.5 L 63.5 234.5 Z M 63.5 216.5 " transform="matrix(1,0,0,1,-62,-104)"/>
+<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 216.5 L 135.5 216.5 L 135.5 234.5 L 63.5 234.5 Z M 63.5 216.5 " transform="matrix(1,0,0,1,-62,-68)"/>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-5" x="19" y="124.5"/>
-  <use xlink:href="#glyph0-6" x="28" y="124.5"/>
-  <use xlink:href="#glyph0-7" x="37" y="124.5"/>
-  <use xlink:href="#glyph0-8" x="46" y="124.5"/>
+  <use xlink:href="#glyph0-5" x="19" y="160.5"/>
+  <use xlink:href="#glyph0-6" x="28" y="160.5"/>
+  <use xlink:href="#glyph0-7" x="37" y="160.5"/>
+  <use xlink:href="#glyph0-8" x="46" y="160.5"/>
 </g>
-<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 180.5 L 135.5 180.5 L 135.5 198.5 L 63.5 198.5 Z M 63.5 180.5 " transform="matrix(1,0,0,1,-62,-104)"/>
+<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 180.5 L 135.5 180.5 L 135.5 198.5 L 63.5 198.5 Z M 63.5 180.5 " transform="matrix(1,0,0,1,-62,-68)"/>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-9" x="19" y="88.5"/>
-  <use xlink:href="#glyph0-10" x="28" y="88.5"/>
-  <use xlink:href="#glyph0-11" x="37" y="88.5"/>
-  <use xlink:href="#glyph0-12" x="46" y="88.5"/>
+  <use xlink:href="#glyph0-9" x="19" y="124.5"/>
+  <use xlink:href="#glyph0-10" x="28" y="124.5"/>
+  <use xlink:href="#glyph0-11" x="37" y="124.5"/>
+  <use xlink:href="#glyph0-12" x="46" y="124.5"/>
 </g>
-<path style="fill-rule:nonzero;fill:rgb(100%,100%,100%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 180.5 L 270.5 180.5 L 270.5 198.5 L 198.5 198.5 Z M 198.5 180.5 " transform="matrix(1,0,0,1,-62,-104)"/>
+<path style="fill-rule:nonzero;fill:rgb(94.940186%,97.225952%,99.389648%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 180.5 L 270.5 180.5 L 270.5 198.5 L 198.5 198.5 Z M 198.5 180.5 " transform="matrix(1,0,0,1,-62,-68)"/>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-5" x="149.5" y="88.5"/>
-  <use xlink:href="#glyph0-6" x="158.5" y="88.5"/>
-  <use xlink:href="#glyph0-13" x="167.5" y="88.5"/>
-  <use xlink:href="#glyph0-14" x="176.5" y="88.5"/>
-  <use xlink:href="#glyph0-12" x="185.5" y="88.5"/>
+  <use xlink:href="#glyph0-13" x="148.105" y="124.5"/>
+  <use xlink:href="#glyph0-14" x="151.714" y="124.5"/>
+  <use xlink:href="#glyph0-15" x="154.711" y="124.5"/>
+  <use xlink:href="#glyph0-14" x="159.292" y="124.5"/>
+  <use xlink:href="#glyph0-13" x="162.289" y="124.5"/>
+  <use xlink:href="#glyph0-14" x="165.898" y="124.5"/>
+  <use xlink:href="#glyph0-5" x="168.895" y="124.5"/>
+  <use xlink:href="#glyph0-6" x="177.895" y="124.5"/>
+  <use xlink:href="#glyph0-12" x="186.895" y="124.5"/>
 </g>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph1-1" x="17.20025" y="22"/>
-  <use xlink:href="#glyph1-2" x="22.53725" y="22"/>
+  <use xlink:href="#glyph1-1" x="17.20025" y="58"/>
+  <use xlink:href="#glyph1-2" x="22.53725" y="58"/>
 </g>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-1" x="24.86825" y="22"/>
-  <use xlink:href="#glyph0-15" x="33.86825" y="22"/>
-  <use xlink:href="#glyph0-16" x="36.86525" y="22"/>
-  <use xlink:href="#glyph0-17" x="42.31025" y="22"/>
-  <use xlink:href="#glyph0-18" x="47.71025" y="22"/>
-  <use xlink:href="#glyph0-19" x="50.08625" y="22"/>
-  <use xlink:href="#glyph0-20" x="55.48625" y="22"/>
-  <use xlink:href="#glyph0-21" x="57.86225" y="22"/>
+  <use xlink:href="#glyph0-1" x="24.86825" y="58"/>
+  <use xlink:href="#glyph0-16" x="33.86825" y="58"/>
+  <use xlink:href="#glyph0-17" x="36.86525" y="58"/>
+  <use xlink:href="#glyph0-18" x="42.31025" y="58"/>
+  <use xlink:href="#glyph0-19" x="47.71025" y="58"/>
+  <use xlink:href="#glyph0-20" x="50.08625" y="58"/>
+  <use xlink:href="#glyph0-21" x="55.48625" y="58"/>
+  <use xlink:href="#glyph0-14" x="57.86225" y="58"/>
 </g>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-22" x="16.89875" y="35"/>
-  <use xlink:href="#glyph0-23" x="22.17275" y="35"/>
-  <use xlink:href="#glyph0-2" x="25.16975" y="35"/>
-  <use xlink:href="#glyph0-15" x="34.16975" y="35"/>
-  <use xlink:href="#glyph0-16" x="37.16675" y="35"/>
-  <use xlink:href="#glyph0-17" x="42.61175" y="35"/>
-  <use xlink:href="#glyph0-18" x="48.01175" y="35"/>
-  <use xlink:href="#glyph0-19" x="50.38775" y="35"/>
-  <use xlink:href="#glyph0-20" x="55.78775" y="35"/>
-  <use xlink:href="#glyph0-21" x="58.16375" y="35"/>
+  <use xlink:href="#glyph0-22" x="16.89875" y="71"/>
+  <use xlink:href="#glyph0-23" x="22.17275" y="71"/>
+  <use xlink:href="#glyph0-2" x="25.16975" y="71"/>
+  <use xlink:href="#glyph0-16" x="34.16975" y="71"/>
+  <use xlink:href="#glyph0-17" x="37.16675" y="71"/>
+  <use xlink:href="#glyph0-18" x="42.61175" y="71"/>
+  <use xlink:href="#glyph0-19" x="48.01175" y="71"/>
+  <use xlink:href="#glyph0-20" x="50.38775" y="71"/>
+  <use xlink:href="#glyph0-21" x="55.78775" y="71"/>
+  <use xlink:href="#glyph0-14" x="58.16375" y="71"/>
 </g>
 <g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
-  <use xlink:href="#glyph0-24" x="33.03125" y="48"/>
+  <use xlink:href="#glyph0-24" x="33.03125" y="84"/>
 </g>
 <g clip-path="url(#clip2)" clip-rule="nonzero">
-<use xlink:href="#image8" transform="matrix(0.0989011,0,0,0.101786,1,148.000002)"/>
-</g>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 252.5 L 135.5 252.5 L 135.5 309.5 L 63.5 309.5 Z M 63.5 252.5 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 252 L 234 243.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 235.898438 L 234 243.898438 M 231 243.898438 L 234 235.898438 L 237 243.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 216 L 234 207.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 199.898438 L 234 207.898438 M 231 207.898438 L 234 199.898438 L 237 207.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 180 L 234 171.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 163.898438 L 234 171.898438 M 231 171.898438 L 234 163.898438 L 237 171.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 252 L 99 243.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 235.898438 L 99 243.898438 M 96 243.898438 L 99 235.898438 L 102 243.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 216 L 99 207.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 199.898438 L 99 207.898438 M 96 207.898438 L 99 199.898438 L 102 207.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99.089844 180 L 99.191406 169.898438 " transform="matrix(1,0,0,1,-62,-104)"/>
-<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99.269531 161.898438 L 99.191406 169.898438 M 96.191406 169.871094 L 99.269531 161.898438 L 102.191406 169.929688 " transform="matrix(1,0,0,1,-62,-104)"/>
+<use xlink:href="#image8" transform="matrix(0.0989011,0,0,0.101786,1,184.000002)"/>
+</g>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 63.5 252.5 L 135.5 252.5 L 135.5 309.5 L 63.5 309.5 Z M 63.5 252.5 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 252 L 234 243.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 235.898438 L 234 243.898438 M 231 243.898438 L 234 235.898438 L 237 243.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 216 L 234 207.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 199.898438 L 234 207.898438 M 231 207.898438 L 234 199.898438 L 237 207.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 144 L 234 135.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 127.898438 L 234 135.898438 M 231 135.898438 L 234 127.898438 L 237 135.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 252 L 99 243.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 235.898438 L 99 243.898438 M 96 243.898438 L 99 235.898438 L 102 243.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 216 L 99 207.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99 199.898438 L 99 207.898438 M 96 207.898438 L 99 199.898438 L 102 207.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99.089844 180 L 99.191406 169.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 99.269531 161.898438 L 99.191406 169.898438 M 96.191406 169.871094 L 99.269531 161.898438 L 102.191406 169.929688 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill-rule:nonzero;fill:rgb(94.940186%,97.225952%,99.389648%);fill-opacity:1;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 198.5 144.5 L 270.5 144.5 L 270.5 162.5 L 198.5 162.5 Z M 198.5 144.5 " transform="matrix(1,0,0,1,-62,-68)"/>
+<g style="fill:rgb(0%,0%,0%);fill-opacity:1;">
+  <use xlink:href="#glyph0-5" x="149.5" y="88.5"/>
+  <use xlink:href="#glyph0-6" x="158.5" y="88.5"/>
+  <use xlink:href="#glyph0-25" x="167.5" y="88.5"/>
+  <use xlink:href="#glyph0-26" x="176.5" y="88.5"/>
+  <use xlink:href="#glyph0-12" x="185.5" y="88.5"/>
+</g>
+<path style="fill:none;stroke-width:1;stroke-linecap:round;stroke-linejoin:round;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 180 L 234 171.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
+<path style="fill:none;stroke-width:1;stroke-linecap:butt;stroke-linejoin:miter;stroke:rgb(0%,0%,0%);stroke-opacity:1;stroke-miterlimit:10;" d="M 234 163.898438 L 234 171.898438 M 231 171.898438 L 234 163.898438 L 237 171.898438 " transform="matrix(1,0,0,1,-62,-68)"/>
 </g>
 </svg>