How to visualize the mask you generate，like the picture in your paper?And question about got10k

[sherlockedlee]

Thanks for your brilliant work.I wonder how to draw picture like this in your paper,can you give me some advise? Thank you very much. And also, can you give the parameter you use to test on got-10k? The result I get is inferior to yours.

[alanlukezic]

Mask can be drawn on image using the following code (m is a binary uint8 numpy array, img is a 3-channel BGR image, alpha is a parameter (0-1) controlling degree of a mask transparency and color is a 3-element array denoting a BGR mask color):

img_r = img[:, :, 2] img_g = img[:, :, 1] img_b = img[:, :, 0] img_r[m > 0] = (1 - alpha) img_r[m > 0] + alpha color[2] img_g[m > 0] = (1 - alpha) img_g[m > 0] + alpha color[1] img_b[m > 0] = (1 - alpha) img_b[m > 0] + alpha color[0]

draw contour around mask

M = (m > 0).astype(np.uint8) if cv2.version[-5] == '4': contours, _ = cv2.findContours(M, cv2.RETR_LIST, cv2.CHAIN_APPROXNONE) else: , contours, _ = cv2.findContours(M, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE) cv2.drawContours(img, contours, -1, color, line_width)

[sherlockedlee]

Mask can be drawn on image using the following code (m is a binary uint8 numpy array, img is a 3-channel BGR image, alpha is a parameter (0-1) controlling degree of a mask transparency and color is a 3-element array denoting a BGR mask color):

img_r = img[:, :, 2] img_g = img[:, :, 1] img_b = img[:, :, 0] img_r[m > 0] = (1 - alpha) img_r[m > 0] + alpha color[2] img_g[m > 0] = (1 - alpha) img_g[m > 0] + alpha color[1] img_b[m > 0] = (1 - alpha) img_b[m > 0] + alpha color[0]

draw contour around mask

M = (m > 0).astype(np.uint8) if cv2.version[-5] == '4': contours, _ = cv2.findContours(M, cv2.RETR_LIST, cv2.CHAIN_APPROXNONE) else: , contours, _ = cv2.findContours(M, cv2.RETR_LIST, cv2.CHAIN_APPROX_NONE) cv2.drawContours(img, contours, -1, color, line_width)

Thank you very much for your patience! And do you use the same parameter in segm_default.py to test on got10k? Mine is inferior to yours. @alanlukezic

[alanlukezic]

Yes, the parameters are the same for experiments on all datasets. We have found out that there might be some differences in performance according to the hardware and library versions.

[sherlockedlee]

Yes, the parameters are the same for experiments on all datasets. We have found out that there might be some differences in performance according to the hardware and library versions.

Thanks alot.