Whether I can use this code on windows7/10?

[xdsonglinliu]

Wonderful work. My platform is Windows 10+caffe+py-R-FCN and it works well.
In order to use soft-nms, I added the code in nms_wrapper.py, cpu_nms.py and config.py following your guide. Then compile the lib folder.
But when I test the caffemodel using soft_nms, I gotted error below.
'ImportError: cannot import name cpu_soft_nms'
Is there anyone has test soft-nms successfully on Windows?

[niuniu111]

Hello, I was also executed on windows. I haven't succeeded yet. If you solve it, can you give me some experience? Thank you.

[bharatsingh430]

you can use the python code

def rescore(overlap, scores, thresh, type='gaussian'):
assert overlap.shape[0] == scores.shape[0]
if type == 'linear':
inds = np.where(overlap >= thresh)[0]
scores[inds] = scores[inds] * (1 - overlap[inds])
else:
scores = scores * np.exp(- overlap**2 / thresh)
return scores

def soft_nms(dets, thresh, max_dets):
if dets.shape[0] == 0:
return np.zeros((0, 5))
x1 = dets[:, 0]
y1 = dets[:, 1]
x2 = dets[:, 2]
y2 = dets[:, 3]
scores = dets[:, 4]

areas = (x2 - x1 + 1) * (y2 - y1 + 1)
order = scores.argsort()[::-1]
scores = scores[order]

if max_dets == -1:
    max_dets = order.size

keep = np.zeros(max_dets, dtype=np.intp)
keep_cnt = 0

while order.size > 0 and keep_cnt < max_dets:
    i = order[0]
    dets[i, 4] = scores[0]
    xx1 = np.maximum(x1[i], x1[order[1:]])
    yy1 = np.maximum(y1[i], y1[order[1:]])
    xx2 = np.minimum(x2[i], x2[order[1:]])
    yy2 = np.minimum(y2[i], y2[order[1:]])

    w = np.maximum(0.0, xx2 - xx1 + 1)
    h = np.maximum(0.0, yy2 - yy1 + 1)
    inter = w * h
    ovr = inter / (areas[i] + areas[order[1:]] - inter)

    order = order[1:]
    scores = rescore(ovr, scores[1:], thresh)

    tmp = scores.argsort()[::-1]
    order = order[tmp]
    scores = scores[tmp]

    keep[keep_cnt] = i
    keep_cnt += 1

keep = keep[:keep_cnt]
dets = dets[keep, :]
return dets

[niuniu111]

Thank you for your reply. I have another question. Max_dets = max (dets)? @bharatsingh430

[bharatsingh430]

Should be length, check the msra repo, I pasted from there

On Mon, May 21, 2018 at 9:28 AM niuniu111 ***@***.***> wrote: Thank you for your reply. I have another question. Max_dets = max (dets)? @bharatsingh430 <https://github.com/bharatsingh430> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#8 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFpAgq3BAuDr7bpNCaDlIPLQ1D-dHD-Qks5t0sERgaJpZM4OnizH> .

-- Regards, Bharat

[lofyol]

Thank you for your reply. I have another question. Max_dets = max (dets)? @bharatsingh430

请问你解决了吗？Have you solved it yet?