2_object_detection

English|中文

Detection samples

Directory structure and description

This catalog contains a variety of test samples for users' reference. The directory structure and specific instructions are as follows.

• yolov3 series samples

  Sample name	Sample description	Characteristic analysis	support chip
  YOLOV3_coco_detection_picture	Picture detection	The input and output are all JPG images, using the acllite public library, the model model is the yolov3-caffe network, and the data set is the coco data set	Ascend310
  YOLOV3_coco_detection_picture _with_postprocess_op	Picture detection	The input and output are all JPG images, using the acllite public library, the model model is the yolov3-tf network, the data set is the coco data set, and the post-processing uses a single operator model	Ascend310
  YOLOV3_mask_detection_picture	Picture detection	The input and output are all JPG images, using the acllite public library, the model model is the yolov3-tf network, and the data set is the mask recognition data set	Ascend310
  YOLOV3_mask_detection_video	Video detection	The input is an mp4 file, the output is a presenter display, using the acllite public library, the model is the yolov3-tf network, and the data set is the mask recognition data set	Ascend310
  coco_detection_rtsp	Video detection	The input is rtsp video stream or H264 file, the output is terminal interface screen display, using acllite public library, model is yolov3-caffe network, data set is coco data set	Ascend310

• yolov4 series samples

  Sample name	Sample description	Characteristic analysis	support chip
  YOLOV4_coco_detection_car_picture	Picture detection	The input and output are all JPG pictures, and the model is yolov4-onnx network. Output lane line and vehicle detection information	Ascend310
  YOLOV4_coco_detection_car_video	Video detection	The input and output are all video files, and the model is yolov4-onnx network. Output lane line and vehicle detection information	Ascend310
  YOLOV4_coco_detection_picture	Video detection	The input and output are all video files, and the model is yolov4-onnx network. Output all test results of COCO data set	Ascend310