Towards-Realtime-MOT
Towards-Realtime-MOT copied to clipboard
Zhongdao
Joint Detection and Embedding for fast multi-object tracking
python demo.py --input-video /mnt/d/object_tracking/crowd.mp4 --weights /mnt/d/object_tracking/yolov4-person_best.weights --output-format video --output-root /mnt/d/object_tracking/tracking_video Namespace(cfg='cfg/yolov3_1088x608.cfg', conf_thres=0.5, input_video='/mnt/d/object_tracking/crowd.mp4', iou_thres=0.5, min_box_area=200, nms_thres=0.4, output_format='video', output_root='/mnt/d/object_tracking/tracking_video', track_buffer=30, weights='/mnt/d/object_tracking/yolov4-person_best.weights') 2020-09-09 14:28:49 [INFO]: Starting tracking... Lenth of the video: 341...
Hi @Zhongdao Excellent work mate. But I am only getting 7 FPS on Windows using the JDE-864x480 weights. How to speed things up?
why the init of kalman filter's std is: std = [ 2 * self._std_weight_position * measurement[3], 2 * self._std_weight_position * measurement[3], 1e-2, 2 * self._std_weight_position * measurement[3], 10 * self._std_weight_velocity...
(Env) C:\ds-projects\Experiments\Towards-Realtime-MOT>python demo.py --input-video C:\ds-projects\Experiments\Towards-Realtime-MOT\videos\MOT16-03.mp4 --weights C:\ds-projects\Experiments\Towards-Realtime-MOT\model\all_dla34.pth --output-root C:\ds-projects\Experiments\Towards-Realtime-MOT Namespace(cfg='cfg/yolov3_1088x608.cfg', conf_thres=0.5, input_video='C:\\ds-projects\\Experiments\\Towards-Realtime-MOT\\videos\\MOT16-03.mp4', iou_thres=0.5, min_box_area=200, nms_thres=0.4, output_format='video', output_root='C:\\ds-projects\\Experiments\\Towards-Realtime-MOT', track_buffer=30, weights='C:\\ds-projects\\Experiments\\Towards-Realtime-MOT\\model\\all_dla34.pth') 2020-08-14 18:09:42 [INFO]: Starting tracking... Lenth of the video: 1500 frames...
1. I have changed DP to DPP, which leads to about 18% speedup during training when using a single node with multi GPUs. (Note that the training script should be...
Just found 'lap' is required after installation. Add 'lap' in requirements to smooth the installation process.
Can the program output the path of pedestrian and display in the output video?
Hello, I ve managed to run the demo, but it seems that it can not track cars. Actually it seems that the demo can only track persons. Is that intented?...
