tps 前反向对齐

ppocr/modeling/transforms/tps.py

@@ -203,9 +203,9 @@ class GridGenerator(nn.Layer):
     def build_C_paddle(self):
         """ Return coordinates of fiducial points in I_r; C """
         F = self.F
-        ctrl_pts_x = paddle.linspace(-1.0, 1.0, int(F / 2))
-        ctrl_pts_y_top = -1 * paddle.ones([int(F / 2)])
-        ctrl_pts_y_bottom = paddle.ones([int(F / 2)])
+        ctrl_pts_x = paddle.linspace(-1.0, 1.0, int(F / 2), dtype='float64')
+        ctrl_pts_y_top = -1 * paddle.ones([int(F / 2)], dtype='float64')
+        ctrl_pts_y_bottom = paddle.ones([int(F / 2)], dtype='float64')
         ctrl_pts_top = paddle.stack([ctrl_pts_x, ctrl_pts_y_top], axis=1)
         ctrl_pts_bottom = paddle.stack([ctrl_pts_x, ctrl_pts_y_bottom], axis=1)
         C = paddle.concat([ctrl_pts_top, ctrl_pts_bottom], axis=0)
@@ -213,12 +213,16 @@ class GridGenerator(nn.Layer):
 
     def build_P_paddle(self, I_r_size):
         I_r_height, I_r_width = I_r_size
-        I_r_grid_x = (
-            paddle.arange(-I_r_width, I_r_width, 2).astype('float32') + 1.0
-        ) / I_r_width  # self.I_r_width
-        I_r_grid_y = (
-            paddle.arange(-I_r_height, I_r_height, 2).astype('float32') + 1.0
-        ) / I_r_height  # self.I_r_height
+        I_r_grid_x = paddle.divide(
+            paddle.arange(
+                -I_r_width, I_r_width, 2, dtype='float64') + 1.0,
+            paddle.to_tensor(
+                I_r_width, dtype='float64'))  # / 2.2128224363981985e-08
+        I_r_grid_y = paddle.divide(
+            paddle.arange(
+                -I_r_height, I_r_height, 2, dtype='float64') + 1.0,
+            paddle.to_tensor(
+                I_r_height, dtype='float64'))  # self.I_r_height
         # P: self.I_r_width x self.I_r_height x 2
         P = paddle.stack(paddle.meshgrid(I_r_grid_x, I_r_grid_y), axis=2)
         P = paddle.transpose(P, perm=[1, 0, 2])
@@ -228,7 +232,7 @@ class GridGenerator(nn.Layer):
     def build_inv_delta_C_paddle(self, C):
         """ Return inv_delta_C which is needed to calculate T """
         F = self.F
-        hat_C = paddle.zeros((F, F), dtype='float32')  # F x F
+        hat_C = paddle.zeros((F, F), dtype='float64')  # F x F
         for i in range(0, F):
             for j in range(i, F):
                 if i == j:
@@ -241,13 +245,21 @@ class GridGenerator(nn.Layer):
         delta_C = paddle.concat(  # F+3 x F+3
             [
                 paddle.concat(
-                    [paddle.ones((F, 1)), C, hat_C], axis=1),  # F x F+3
+                    [paddle.ones(
+                        (F, 1), dtype='float64'), C, hat_C], axis=1),  # F x F+3
                 paddle.concat(
-                    [paddle.zeros((2, 3)), paddle.transpose(
-                        C, perm=[1, 0])],
+                    [
+                        paddle.zeros(
+                            (2, 3), dtype='float64'), paddle.transpose(
+                                C, perm=[1, 0])
+                    ],
                     axis=1),  # 2 x F+3
                 paddle.concat(
-                    [paddle.zeros((1, 3)), paddle.ones((1, F))],
+                    [
+                        paddle.zeros(
+                            (1, 3), dtype='float64'), paddle.ones(
+                                (1, F), dtype='float64')
+                    ],
                     axis=1)  # 1 x F+3
             ],
             axis=0)
@@ -268,7 +280,9 @@ class GridGenerator(nn.Layer):
         # rbf: n x F
         rbf = paddle.multiply(
             paddle.square(rbf_norm), paddle.log(rbf_norm + eps))
-        P_hat = paddle.concat([paddle.ones((n, 1)), P, rbf], axis=1)
+        P_hat = paddle.concat(
+            [paddle.ones(
+                (n, 1), dtype='float64'), P, rbf], axis=1)
         return P_hat  # n x F+3
 
     def get_expand_tensor(self, batch_C_prime):
@@ -296,3 +310,34 @@ class TPS(nn.Layer):
             [-1, image.shape[2], image.shape[3], 2])
         batch_I_r = F.grid_sample(x=image, grid=batch_P_prime)
         return batch_I_r
+
+
+if __name__ == '__main__':
+    import paddle
+    from ppocr.utils.save_load import load_dygraph_pretrain
+    from ppocr.utils.logging import get_logger
+
+    np.random.seed(0)
+    img = np.random.random((1, 3, 32, 100)).astype(np.float32)
+    x = paddle.to_tensor(img)
+
+    batch_P_prime = np.load(
+        '/Users/zhoujun20/Desktop/code/static/PaddleOCR/output/tps1.npy')
+    batch_P_prime = paddle.to_tensor(batch_P_prime)
+
+    model = TPS(in_channels=3, num_fiducial=20, loc_lr=0.1, model_name='small')
+    load_dygraph_pretrain(
+        model,
+        get_logger(),
+        '/Users/zhoujun20/Desktop/code/static/PaddleOCR/output/tps',
+        load_static_weights=True)
+
+    model.eval()
+    y = model(x)
+    print(y.shape)
+
+    static = np.load(
+        '/Users/zhoujun20/Desktop/code/static/PaddleOCR/output/tps.npy')
+    print(static.dtype, y.numpy().dtype, static.shape, y.numpy().shape)
+    diff = y.numpy() - static
+    print(diff.max())