hpc203 / yolov7-detect-face-onnxrun-cpp-py

使用ONNXRuntime部署YOLOV7人脸+关键点检测,包含C++和Python两个版本的程序
50 stars 7 forks source link

hi, the postprocessing seems have some problem? #1

Closed lucasjinreal closed 1 year ago

lucasjinreal commented 2 years ago

Most of images looks normal, except this image:

image

as you can see at first glance, some face are more like horizontal rect rather than normally vertical rect,

what could be the reason?

KangChou commented 1 year ago

Traceback (most recent call last): File "D:\CV2022\mymediapipe\yolov7\exe_yolov7\main.py", line 158, in <module> dstimg = YOLOv7_face_detector.draw_detections(srcimg, boxes, scores, kpts) File "D:\CV2022\mymediapipe\yolov7\exe_yolov7\main.py", line 123, in draw_detections x, y, w, h = box.astype(int) ValueError: not enough values to unpack (expected 4, got 1)