opencv/opencv_contrib
1
0
Fork 0
mirror of https://github.com/opencv/opencv_contrib.git synced 2025-10-23 00:49:38 +08:00
Code Issues Releases Wiki Activity

Merge pull request #3821 from asmorkalov:as/new_dnn_caffe

Browse Source 
Do not use forward to layer feature in WeChat to support the new dnn engine
This commit is contained in:
Alexander Smorkalov
2024-11-06 17:32:30 +03:00
committed by GitHub
parent bd3b734731 aaad40afbc
commit fe10e50fb7
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
modules/wechat_qrcode/src/detector/ssd_detector.cpp
View File

@@ -24,7 +24,7 @@ vector<Mat> SSDDetector::forward(Mat img, const int target_width, const int targ
false, false); false, false);
net_.setInput(input, "data"); net_.setInput(input, "data");
auto prob = net_.forward("detection_output"); auto prob = net_.forward();
vector<Mat> point_list; vector<Mat> point_list;
// the shape is (1,1,100,7)=>(batch,channel,count,dim) // the shape is (1,1,100,7)=>(batch,channel,count,dim)
for (int row = 0; row < prob.size[2]; row++) { for (int row = 0; row < prob.size[2]; row++) {