How can we use pre-trained models of scrfd in C++ on Ubuntu? Can someone please provide a minimum working and self-contained example that reads an image, detects faces using scrfd and outputs the results?
iecba09b1#
There is scrfd.cpp in ncnn/examples folder
p8h8hvxi2#
想问Insightface这边是用stack的方式做anchor_center
与ncnn/examples/scrfd.cpp里却是用generate_anchors(int base_size, const ncnn::Mat& ratios, const ncnn::Mat& scales)来做anchor而且这段是参考gen_single_level_base_anchors()改写而来
想问两者之间的差异?
2条答案
按热度按时间iecba09b1#
There is scrfd.cpp in ncnn/examples folder
p8h8hvxi2#
想问Insightface这边是用stack的方式做anchor_center
与ncnn/examples/scrfd.cpp里却是用
generate_anchors(int base_size, const ncnn::Mat& ratios, const ncnn::Mat& scales)来做anchor
而且这段是参考gen_single_level_base_anchors()改写而来
想问两者之间的差异?