About Region Confusion Mechanism

[lao-ling-jie]

Hi！it's great work! But I have some questions about the region confusion mechanism in the code (below). What's the mean of the parameter pro and RAN，and why pro should >=5, I can't find this information in your paper.

pro = 5
    if pro >= 5:          
        tmpx = []
        tmpy = []
        count_x = 0
        count_y = 0
        k = 2
        RAN = 2
        for i in range(crop[1] * crop[0]):
            tmpx.append(images[i])
            count_x += 1
            if len(tmpx) >= k:
                tmp = tmpx[count_x - RAN:count_x]
                random.shuffle(tmp)
                tmpx[count_x - RAN:count_x] = tmp
            if count_x == crop[0]:
                tmpy.append(tmpx)
                count_x = 0
                count_y += 1
                tmpx = []
            if len(tmpy) >= k:
                tmp2 = tmpy[count_y - RAN:count_y]
                random.shuffle(tmp2)
                tmpy[count_y - RAN:count_y] = tmp2
        random_im = []
        for line in tmpy:
            random_im.extend(line)

[XA-kirino]

Hi！it's great work! But I have some questions about the region confusion mechanism in the code (below). What's the mean of the parameter pro and RAN，and why pro should >=5, I can't find this information in your paper.

pro = 5
    if pro >= 5:          
        tmpx = []
        tmpy = []
        count_x = 0
        count_y = 0
        k = 2
        RAN = 2
        for i in range(crop[1] * crop[0]):
            tmpx.append(images[i])
            count_x += 1
            if len(tmpx) >= k:
                tmp = tmpx[count_x - RAN:count_x]
                random.shuffle(tmp)
                tmpx[count_x - RAN:count_x] = tmp
            if count_x == crop[0]:
                tmpy.append(tmpx)
                count_x = 0
                count_y += 1
                tmpx = []
            if len(tmpy) >= k:
                tmp2 = tmpy[count_y - RAN:count_y]
                random.shuffle(tmp2)
                tmpy[count_y - RAN:count_y] = tmp2
        random_im = []
        for line in tmpy:
            random_im.extend(line)

pro might be abbr of probability and only defined once here, maybe will be modified to swap with a prob.
RAN is a param defining neighbor range, of the numbers of rows or columns to shuffle.

[lao-ling-jie]

Hi！it's great work! But I have some questions about the region confusion mechanism in the code (below). What's the mean of the parameter pro and RAN，and why pro should >=5, I can't find this information in your paper.

pro = 5
    if pro >= 5:          
        tmpx = []
        tmpy = []
        count_x = 0
        count_y = 0
        k = 2
        RAN = 2
        for i in range(crop[1] * crop[0]):
            tmpx.append(images[i])
            count_x += 1
            if len(tmpx) >= k:
                tmp = tmpx[count_x - RAN:count_x]
                random.shuffle(tmp)
                tmpx[count_x - RAN:count_x] = tmp
            if count_x == crop[0]:
                tmpy.append(tmpx)
                count_x = 0
                count_y += 1
                tmpx = []
            if len(tmpy) >= k:
                tmp2 = tmpy[count_y - RAN:count_y]
                random.shuffle(tmp2)
                tmpy[count_y - RAN:count_y] = tmp2
        random_im = []
        for line in tmpy:
            random_im.extend(line)

pro might be abbr of probability and only defined once here, maybe will be modified to swap with a prob.
RAN is a param defining neighbor range, of the numbers of rows or columns to shuffle.

Thanks for your reply! So pro in this code might be not useful?

[XA-kirino]

Hi！it's great work! But I have some questions about the region confusion mechanism in the code (below). What's the mean of the parameter pro and RAN，and why pro should >=5, I can't find this information in your paper.

pro = 5
    if pro >= 5:          
        tmpx = []
        tmpy = []
        count_x = 0
        count_y = 0
        k = 2
        RAN = 2
        for i in range(crop[1] * crop[0]):
            tmpx.append(images[i])
            count_x += 1
            if len(tmpx) >= k:
                tmp = tmpx[count_x - RAN:count_x]
                random.shuffle(tmp)
                tmpx[count_x - RAN:count_x] = tmp
            if count_x == crop[0]:
                tmpy.append(tmpx)
                count_x = 0
                count_y += 1
                tmpx = []
            if len(tmpy) >= k:
                tmp2 = tmpy[count_y - RAN:count_y]
                random.shuffle(tmp2)
                tmpy[count_y - RAN:count_y] = tmp2
        random_im = []
        for line in tmpy:
            random_im.extend(line)

pro might be abbr of probability and only defined once here, maybe will be modified to swap with a prob.
RAN is a param defining neighbor range, of the numbers of rows or columns to shuffle.

Thanks for your reply! So pro in this code might be not useful?

Yes

[lao-ling-jie]

Hi！it's great work! But I have some questions about the region confusion mechanism in the code (below). What's the mean of the parameter pro and RAN，and why pro should >=5, I can't find this information in your paper.

pro = 5
    if pro >= 5:          
        tmpx = []
        tmpy = []
        count_x = 0
        count_y = 0
        k = 2
        RAN = 2
        for i in range(crop[1] * crop[0]):
            tmpx.append(images[i])
            count_x += 1
            if len(tmpx) >= k:
                tmp = tmpx[count_x - RAN:count_x]
                random.shuffle(tmp)
                tmpx[count_x - RAN:count_x] = tmp
            if count_x == crop[0]:
                tmpy.append(tmpx)
                count_x = 0
                count_y += 1
                tmpx = []
            if len(tmpy) >= k:
                tmp2 = tmpy[count_y - RAN:count_y]
                random.shuffle(tmp2)
                tmpy[count_y - RAN:count_y] = tmp2
        random_im = []
        for line in tmpy:
            random_im.extend(line)

pro might be abbr of probability and only defined once here, maybe will be modified to swap with a prob.
RAN is a param defining neighbor range, of the numbers of rows or columns to shuffle.

Thanks for your reply! So pro in this code might be not useful?

Yes

Thank you!