Make roi_perspective_transform op return mask and transform matrix (#18371)

* modify roi_perspective_transform_op to output mask and transform matrix

* modify comment

* modify comment

* modify API.spec

* update API.spec

* remove no use header, test=develop

* resolve conflict

paddle/fluid/API.spec

@@ -353,7 +353,7 @@ paddle.fluid.layers.rpn_target_assign (ArgSpec(args=['bbox_pred', 'cls_logits',
 paddle.fluid.layers.retinanet_target_assign (ArgSpec(args=['bbox_pred', 'cls_logits', 'anchor_box', 'anchor_var', 'gt_boxes', 'gt_labels', 'is_crowd', 'im_info', 'num_classes', 'positive_overlap', 'negative_overlap'], varargs=None, keywords=None, defaults=(1, 0.5, 0.4)), ('document', 'fa1d1c9d5e0111684c0db705f86a2595'))
 paddle.fluid.layers.sigmoid_focal_loss (ArgSpec(args=['x', 'label', 'fg_num', 'gamma', 'alpha'], varargs=None, keywords=None, defaults=(2, 0.25)), ('document', 'aeac6aae100173b3fc7f102cf3023a3d'))
 paddle.fluid.layers.anchor_generator (ArgSpec(args=['input', 'anchor_sizes', 'aspect_ratios', 'variance', 'stride', 'offset', 'name'], varargs=None, keywords=None, defaults=(None, None, [0.1, 0.1, 0.2, 0.2], None, 0.5, None)), ('document', '0aaacaf9858b8270a8ab5b0aacdd94b7'))
-paddle.fluid.layers.roi_perspective_transform (ArgSpec(args=['input', 'rois', 'transformed_height', 'transformed_width', 'spatial_scale'], varargs=None, keywords=None, defaults=(1.0,)), ('document', 'd1ddc75629fedee46f82e631e22c79dc'))
+paddle.fluid.layers.roi_perspective_transform (ArgSpec(args=['input', 'rois', 'transformed_height', 'transformed_width', 'spatial_scale'], varargs=None, keywords=None, defaults=(1.0,)), ('document', '54e3bf70e3bdbd58b3b9b65b3c69a854'))
 paddle.fluid.layers.generate_proposal_labels (ArgSpec(args=['rpn_rois', 'gt_classes', 'is_crowd', 'gt_boxes', 'im_info', 'batch_size_per_im', 'fg_fraction', 'fg_thresh', 'bg_thresh_hi', 'bg_thresh_lo', 'bbox_reg_weights', 'class_nums', 'use_random', 'is_cls_agnostic', 'is_cascade_rcnn'], varargs=None, keywords=None, defaults=(256, 0.25, 0.25, 0.5, 0.0, [0.1, 0.1, 0.2, 0.2], None, True, False, False)), ('document', '69def376b42ef0681d0cc7f53a2dac4b'))
 paddle.fluid.layers.generate_proposals (ArgSpec(args=['scores', 'bbox_deltas', 'im_info', 'anchors', 'variances', 'pre_nms_top_n', 'post_nms_top_n', 'nms_thresh', 'min_size', 'eta', 'name'], varargs=None, keywords=None, defaults=(6000, 1000, 0.5, 0.1, 1.0, None)), ('document', 'b7d707822b6af2a586bce608040235b1'))
 paddle.fluid.layers.generate_mask_labels (ArgSpec(args=['im_info', 'gt_classes', 'is_crowd', 'gt_segms', 'rois', 'labels_int32', 'num_classes', 'resolution'], varargs=None, keywords=None, defaults=None), ('document', 'b319b10ddaf17fb4ddf03518685a17ef'))

paddle/fluid/operators/detection/roi_perspective_transform_op.cc

@@ -243,7 +243,9 @@ class CPUROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
     auto* in = ctx.Input<framework::Tensor>("X");
     auto* rois = ctx.Input<framework::LoDTensor>("ROIs");
     auto* out = ctx.Output<framework::Tensor>("Out");
-
+    auto* mask = ctx.Output<framework::Tensor>("Mask");
+    auto* out_transform_matrix =
+        ctx.Output<framework::Tensor>("TransformMatrix");
     auto transformed_height = ctx.Attr<int>("transformed_height");
     auto transformed_width = ctx.Attr<int>("transformed_width");
     auto spatial_scale = ctx.Attr<float>("spatial_scale");
@@ -255,6 +257,7 @@ class CPUROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
     int rois_num = rois->dims()[0];
 
     const T* input_data = in->data<T>();
+    int* mask_data = mask->mutable_data<int>(ctx.GetPlace());
 
     framework::Tensor roi2image;
     roi2image.Resize({rois_num});
@@ -279,7 +282,8 @@ class CPUROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
       }
       int image_id = roi2image_data[n];
       // Get transform matrix
-      T transform_matrix[9];
+      T* transform_matrix =
+          out_transform_matrix->mutable_data<T>({9}, ctx.GetPlace());
       get_transform_matrix<T>(transformed_width, transformed_height, roi_x,
                               roi_y, transform_matrix);
 
@@ -298,13 +302,19 @@ class CPUROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
                   GT<T>(-0.5, in_h) ||
                   GT<T>(in_h, static_cast<T>(in_height - 0.5))) {
                 output_data[out_index] = 0.0;
+                mask_data[(n * transformed_height + out_h) * transformed_width +
+                          out_w] = 0;
               } else {
                 bilinear_interpolate(input_data, channels, in_width, in_height,
                                      image_id, c, in_w, in_h,
                                      output_data + out_index);
+                mask_data[(n * transformed_height + out_h) * transformed_width +
+                          out_w] = 1;
               }
             } else {
               output_data[out_index] = 0.0;
+              mask_data[(n * transformed_height + out_h) * transformed_width +
+                        out_w] = 0;
             }
           }
         }
@@ -467,7 +477,6 @@ class ROIPerspectiveTransformOp : public framework::OperatorWithKernel {
         "Output(Out) of ROIPerspectiveTransformOp should not be null.");
     auto input_dims = ctx->GetInputDim("X");
     auto rois_dims = ctx->GetInputDim("ROIs");
-
     PADDLE_ENFORCE(input_dims.size() == 4,
                    "The format of input tensor is NCHW.");
     PADDLE_ENFORCE(rois_dims.size() == 2,
@@ -476,7 +485,6 @@ class ROIPerspectiveTransformOp : public framework::OperatorWithKernel {
     PADDLE_ENFORCE(rois_dims[1] == 8,
                    "ROIs should be a 2-D LoDTensor of shape (num_rois, 8)"
                    "given as [[x0, y0, x1, y1, x2, y2, x3, y3], ...].");
-
     int transformed_height = ctx->Attrs().Get<int>("transformed_height");
     int transformed_width = ctx->Attrs().Get<int>("transformed_width");
     float spatial_scale = ctx->Attrs().Get<float>("spatial_scale");
@@ -493,7 +501,18 @@ class ROIPerspectiveTransformOp : public framework::OperatorWithKernel {
                                      static_cast<int64_t>(transformed_width)});
     auto out_dims = framework::make_ddim(out_dims_v);
 
+    std::vector<int64_t> mask_dims_v({rois_dims[0],  // num_rois
+                                      1,             // channels
+                                      static_cast<int64_t>(transformed_height),
+                                      static_cast<int64_t>(transformed_width)});
+    auto mask_dims = framework::make_ddim(mask_dims_v);
+
+    std::vector<int64_t> matrix_dims_v(9);
+    auto matrix_dims = framework::make_ddim(matrix_dims_v);
+
     ctx->SetOutputDim("Out", out_dims);
+    ctx->SetOutputDim("Mask", mask_dims);
+    ctx->SetOutputDim("TransformMatrix", matrix_dims);
     ctx->SetOutputDim("Out2InIdx", out_dims);
     ctx->SetOutputDim("Out2InWeights", out_dims);
     ctx->ShareLoD("ROIs", /*->*/ "Out");
@@ -552,6 +571,16 @@ class ROIPerspectiveTransformOpMaker
         "(Tensor), "
         "The output of ROIPerspectiveTransformOp is a 4-D tensor with shape "
         "(num_rois, channels, transformed_h, transformed_w).");
+    AddOutput("Mask",
+              "(Tensor), "
+              "The output mask of ROIPerspectiveTransformOp is a 4-D tensor "
+              "with shape "
+              "(num_rois, 1, transformed_h, transformed_w).");
+    AddOutput("TransformMatrix",
+              "(Tensor), "
+              "The output transform matrix of ROIPerspectiveTransformOp is a "
+              "1-D tensor with shape "
+              "(9,).");
     AddOutput("Out2InIdx",
               "(Tensor), "
               "An intermediate tensor used to map indexes of input feature map "

paddle/fluid/operators/detection/roi_perspective_transform_op.cu

@@ -278,7 +278,7 @@ __global__ void RoiTransformKernel(
     const float* input_data, const float* rois_data, const int* roi2image_data,
     int num_rois, int in_height, int in_width, int channels,
     int transformed_height, int transformed_width, float spatial_scale,
-    T* output_data, int* out2in_idx, T* out2in_w) {
+    T* output_data, int* out2in_idx, T* out2in_w, int* mask, T* matrix) {
   int output_size =
       num_rois * transformed_height * transformed_width * channels;
 
@@ -303,7 +303,6 @@ __global__ void RoiTransformKernel(
     }
 
     // Get transform matrix
-    T matrix[9];
     get_transform_matrix<T>(transformed_width, transformed_height, roi_x, roi_y,
                             matrix);
 
@@ -317,17 +316,20 @@ __global__ void RoiTransformKernel(
           GT<T>(-0.5, in_h) || GT<T>(in_h, static_cast<T>(in_height - 0.5))) {
         // Skip if source coords is not in input image
         output_data[index] = 0.0;
+        mask[(n * transformed_height + out_h) * transformed_width + out_w] = 0;
       } else {
         // Perform bilinear interpolation
         int in_n = roi2image_data[n];
         bilinear_interpolate<T>(input_data, channels, in_width, in_height, in_n,
                                 c, in_w, in_h, output_data + index, index,
                                 out2in_idx, out2in_w);
+        mask[(n * transformed_height + out_h) * transformed_width + out_w] = 1;
       }
 
     } else {
       // Skip if source coords is not in quad
       output_data[index] = 0.0;
+      mask[(n * transformed_height + out_h) * transformed_width + out_w] = 0;
     }
   }
 }
@@ -341,7 +343,11 @@ class CUDAROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
     auto* out = ctx.Output<framework::Tensor>("Out");
     auto* out2in_idx = ctx.Output<framework::Tensor>("Out2InIdx");
     auto* out2in_w = ctx.Output<framework::Tensor>("Out2InWeights");
+    auto* mask = ctx.Output<framework::Tensor>("Mask");
+    auto* out_transform_matrix =
+        ctx.Output<framework::Tensor>("TransformMatrix");
 
+    int* mask_data = mask->mutable_data<int>(ctx.GetPlace());
     int* out2in_idx_data =
         out2in_idx->mutable_data<int>({out->numel(), 4}, ctx.GetPlace());
     T* out2in_w_data =
@@ -382,10 +388,14 @@ class CUDAROIPerspectiveTransformOpKernel : public framework::OpKernel<T> {
     int block = 512;
     int grid = (out_size + block - 1) / block;
 
+    // Get transform matrix
+    T* matrix = out_transform_matrix->mutable_data<T>({9}, ctx.GetPlace());
+
     RoiTransformKernel<T><<<grid, block, 0, stream>>>(
         input_data, rois_data, roi2image_dev.data<int>(), rois_num, in_height,
         in_width, channels, transformed_height, transformed_width,
-        spatial_scale, output_data, out2in_idx_data, out2in_w_data);
+        spatial_scale, output_data, out2in_idx_data, out2in_w_data, mask_data,
+        matrix);
   }
 };
 

python/paddle/fluid/layers/detection.py

@@ -2099,8 +2099,16 @@ def roi_perspective_transform(input,
         spatial_scale (float): Spatial scale factor to scale ROI coords. Default: 1.0
 
     Returns:
-        Variable: The output of ROIPerspectiveTransformOp which is a 4-D tensor with shape 
-                  (num_rois, channels, transformed_h, transformed_w).
+            tuple: A tuple with three Variables. (out, mask, transform_matrix)
+
+            out: The output of ROIPerspectiveTransformOp which is a 4-D tensor with shape
+            (num_rois, channels, transformed_h, transformed_w).
+
+            mask: The mask of ROIPerspectiveTransformOp which is a 4-D tensor with shape
+            (num_rois, 1, transformed_h, transformed_w).
+
+            transform_matrix: The transform matrix of ROIPerspectiveTransformOp which is
+            a 1-D tensor with shape (9,).
 
     Examples:
         .. code-block:: python
@@ -2109,11 +2117,13 @@ def roi_perspective_transform(input,
 
             x = fluid.layers.data(name='x', shape=[256, 28, 28], dtype='float32')
             rois = fluid.layers.data(name='rois', shape=[8], lod_level=1, dtype='float32')
-            out = fluid.layers.roi_perspective_transform(x, rois, 7, 7, 1.0)
+            out, mask, transform_matrix = fluid.layers.roi_perspective_transform(x, rois, 7, 7, 1.0)
     """
     helper = LayerHelper('roi_perspective_transform', **locals())
     dtype = helper.input_dtype()
     out = helper.create_variable_for_type_inference(dtype)
+    mask = helper.create_variable_for_type_inference(dtype="int32")
+    transform_matrix = helper.create_variable_for_type_inference(dtype)
     out2in_idx = helper.create_variable_for_type_inference(dtype="int32")
     out2in_w = helper.create_variable_for_type_inference(dtype)
     helper.append_op(
@@ -2123,14 +2133,16 @@ def roi_perspective_transform(input,
         outputs={
             "Out": out,
             "Out2InIdx": out2in_idx,
-            "Out2InWeights": out2in_w
+            "Out2InWeights": out2in_w,
+            "Mask": mask,
+            "TransformMatrix": transform_matrix
         },
         attrs={
             "transformed_height": transformed_height,
             "transformed_width": transformed_width,
             "spatial_scale": spatial_scale
         })
-    return out
+    return out, mask, transform_matrix
 
 
 def generate_proposal_labels(rpn_rois,

python/paddle/fluid/tests/unittests/test_roi_perspective_transform_op.py

@@ -198,7 +198,8 @@ def roi_transform(in_data, rois, rois_lod, transformed_height,
             roi2image[j] = i
 
     out = np.zeros([rois_num, channels, transformed_height, transformed_width])
-
+    mask = np.zeros(
+        [rois_num, 1, transformed_height, transformed_width]).astype('int')
     for n in range(rois_num):
         roi_x = []
         roi_y = []
@@ -219,9 +220,11 @@ def roi_transform(in_data, rois, rois_lod, transformed_height,
                                 in_h, -0.5) and lt_e(in_h, in_height - 0.5):
                         out[n][c][out_h][out_w] = bilinear_interpolate(
                             in_data, image_id, c, in_w, in_h)
+                        mask[n][0][out_h][out_w] = 1
                     else:
                         out[n][c][out_h][out_w] = 0.0
-    return out.astype("float32")
+                        mask[n][0][out_h][out_w] = 0
+    return out.astype("float32"), mask, transform_matrix
 
 
 class TestROIPoolOp(OpTest):
@@ -236,10 +239,14 @@ class TestROIPoolOp(OpTest):
             'transformed_height': self.transformed_height,
             'transformed_width': self.transformed_width
         }
-        out = roi_transform(self.x, self.rois, self.rois_lod,
-                            self.transformed_height, self.transformed_width,
-                            self.spatial_scale)
-        self.outputs = {'Out': out}
+        out, mask, transform_matrix = roi_transform(
+            self.x, self.rois, self.rois_lod, self.transformed_height,
+            self.transformed_width, self.spatial_scale)
+        self.outputs = {
+            'Out': out,
+            'Mask': mask,
+            'TransformMatrix': transform_matrix
+        }
 
     def init_test_case(self):
         self.batch_size = 2