cholihao/retinaface-caffe

Retinaface caffe with mnetv2 0.25(and many more)

retinaface-caffe's Issues

error C2065: “anchor”: undeclared identifier

@cholihao
void AnchorGenerator::landmark_pred(const CRect2f anchor, const std::vectorcv::Point2f& delta, std::vectorcv::Point2f& pts) {
float w = anchor[2] - anchor[0] + 1;
float h = anchor[3] - anchor[1] + 1;
float x_ctr = anchor[0] + 0.5 * (w - 1);
float y_ctr = anchor[1] + 0.5 * (h - 1);
pts.resize(delta.size());
for (int i = 0; i < delta.size(); ++i) {
pts[i].x = (delta[i].x*w + x_ctr)ratiow;
pts[i].y = (delta[i].yh + y_ctr)*ratioh;
}
}

about landmark

if (pts) {
			std::vector<cv::Point2f> pts_delta(pts_length);
			for (int p = 0; p < pts_length; ++p) {
			    pts_delta[p].x = pts->channel(0, a*pts_length*2+p*2)[id];
			    pts_delta[p].y = pts->channel(0, a*pts_length*2+p*2+1)[id];
			}
			printf("ready landmark_pred\n");
			landmark_pred(box, pts_delta, res.pts);
			printf("landmark_pred\n");
		    }
``` pts have no channel,  request for your help
I have succeed in running detect face, Do you succeed in predicting landmark? 

How to convert mxnet model to caffe?

Thanks for your work, Could you share your convert code? I try to convert it use mmdnn, but not work, if you not convenient, Could you give me some direction for it. I don't know how to process crop and upsampling layer. thx again.

did you miss uploading detect.h?