aadeshnpn/OSDN

Re-calibrate scores

yasserboutaleb opened this issue · 2 comments

Hello,

Please, the "channel_scores" is the feature vector of the "fc8" layer for a specific channel, then why did you loop over the "categoryid" to compute modified_fc8_score = channel_scores[categoryid] * ( 1 - wscore*ranked_alpha[categoryid] ) ?
Yet, the Len(channel_scores) is not equal to NCHANNELS.

Thank you.

Hi, I've noticed the same issue...It seems that the code has been written with the idea to generalize to more channel but in the end he didn't finish to debug it with more than one channel. In fact, it runs only if NCHANNELS=1.

I hope someone will solve this issue soon.

Thanks @sgnet91 and @Bestyasser . The code was written to generalize to many channels but for now it only works for grey images. I have updated the repo with example and details step to use your own model and data. Hope this helps.