フィルターのクリア

neural network toolbox _faster rcnn_code error

1 回表示 (過去 30 日間)
Mingrun Wang
Mingrun Wang 2018 年 5 月 2 日
回答済み: Mingrun Wang 2018 年 5 月 2 日
have anyone found a coding error in the module of faster_rcnn under the neural network toolbox? the error is in the path "...MATLAB\R2018a\toolbox\vision\vision\+vision\+internal\+cnn\+rpn\selectTrainingSamples.m" the error is located within the line 114 to line 116,where the var 'negIdx' is the negativeIndex, it's also an aaray and it's derived from the var 'negativeIndex' by invoking the function 'find(negativeIndex)'.the var 'nidx' is the random selected index of var 'negIdx'.So should the description of random_index selecting of the var 'negativeIndex'(an array) be written with 'negIdx(nidx)' rather than 'nidx'?

回答 (1 件)

Mingrun Wang
Mingrun Wang 2018 年 5 月 2 日
this error may have no any essential significants to rpn training ,because the number of postive box samples is often too small.

カテゴリ

Help Center および File ExchangeDeep Learning Toolbox についてさらに検索

タグ

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by