getting segmentation points as pixel value

[senemaktas]

getting segmentation points as pixel value

[senemaktas]

I got the following code from Yolov5. Those who want can use it as follows:

import numpy as np

def masks2segments(masks, strategy='largest'):
    # Convert masks(n,160,160) into segments(n,xy)
    segments = []
    for x in masks.int().cpu().numpy().astype('uint8'):
        c = cv2.findContours(x, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)[0]
        if c:
            if strategy == 'concat':  # concatenate all segments
                c = np.concatenate([x.reshape(-1, 2) for x in c])
            elif strategy == 'largest':  # select largest segment
                c = np.array(c[np.array([len(x) for x in c]).argmax()]).reshape(-1, 2)
        else:
            c = np.zeros((0, 2))  # no segments found
        segments.append(c.astype('float32'))
    return segments

def clip_segments(segments, shape):
    # Clip segments (xy1,xy2,...) to image shape (height, width)
    if isinstance(segments, torch.Tensor):  # faster individually
        segments[:, 0].clamp_(0, shape[1])  # x
        segments[:, 1].clamp_(0, shape[0])  # y
    else:  # np.array (faster grouped)
        segments[:, 0] = segments[:, 0].clip(0, shape[1])  # x
        segments[:, 1] = segments[:, 1].clip(0, shape[0])  # y

def scale_segments(img1_shape, segments, img0_shape, ratio_pad=None, normalize=False):
    # Rescale coords (xyxy) from img1_shape to img0_shape
    if ratio_pad is None:  # calculate from img0_shape
        gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1])  # gain  = old / new
        pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2  # wh padding
    else:
        gain = ratio_pad[0][0]
        pad = ratio_pad[1]

    segments[:, 0] -= pad[0]  # x padding
    segments[:, 1] -= pad[1]  # y padding
    segments /= gain
    clip_segments(segments, img0_shape)
    if normalize:
        segments[:, 0] /= img0_shape[1]  # width
        segments[:, 1] /= img0_shape[0]  # height
    return segments

    ******
   # use where you get masks
              masks = process_mask(proto[i], det[:, 6:], det[:, :4], im.shape[2:], upsample=True)  # HWC
              retina_masks = False
              segments = [scale_segments(im0.shape if retina_masks else im.shape[2:], x, im0.shape, normalize=True)
                          for x in reversed(masks2segments(masks))]

              image = cv2.imread(source)  
              h, w, _ = image.shape
              for each_seg in segments:
                  each_converted_seg_point = []
                  for point in each_seg:
                      x, y = point[0], point[1]
                      x1 = x * w
                      y1 = y * h
                      each_converted_seg_point.append([x1, y1])
                      cv2.circle(image, (int(x1), int(y1)), radius=5, color=(0, 0, 255), thickness=-1)

                  pts = np.array(each_converted_seg_point, np.int32)
                  pts = pts.reshape((-1, 1, 2))
                  cv2.polylines(image, [pts], isClosed=True, color=(0, 255, 0), thickness=2)

              cv2.imshow("ddd", image)
              cv2.waitKey(0)