Towards-Realtime-MOT icon indicating copy to clipboard operation
Towards-Realtime-MOT copied to clipboard

Joint Detection and Embedding for fast multi-object tracking

Results 105 Towards-Realtime-MOT issues
Sort by recently updated
recently updated
newest added

Command : `python demo.py --input-video /path/to/input/video.mp4 --weights /path/to/weights/jde_576x320_uncertainty.pt --output-format video --cfg cfg/yolov3_576x320.cfg` Output/Error: Namespace(cfg='cfg/yolov3_576x320.cfg', conf_thres=0.5, input_video='/path/to/input/video.mp4', iou_thres=0.5, min_box_area=200, nms_thres=0.4, output_format='video', output_root='results', track_buffer=30, weights='/path/to/weights/jde_576x320_uncertainty.pt') 2020-05-29 17:42:12 [INFO]: Starting tracking... Lenth of...

https://github.com/Zhongdao/Towards-Realtime-MOT/blob/44907e13b14108d5a12b9dad4eeb63603999e092/tracker/multitracker.py#L269-L310 Why not use appearance affinity to associate unmatched detections in the current frame with unconfirmed tracklets which were created as new tracks in the last frame? Why use different...

https://github.com/Zhongdao/Towards-Realtime-MOT/blob/44907e13b14108d5a12b9dad4eeb63603999e092/tracker/multitracker.py#L336-L337 Line 336 should be `self.lost_stracks = sub_stracks(self.lost_stracks, removed_stracks)` given the context of the code. Or swap Line 336 and Line 337.

# matching.py from scipy.spatial.distance import cdist def embedding_distance(tracks, detections, metric='cosine'): cost_matrix = np.zeros((len(tracks), len(detections)), dtype=np.float) if cost_matrix.size == 0: return cost_matrix det_features = np.asarray([track.curr_feat for track in detections], dtype=np.float) track_features...

hi,I want to implement ship tracking How to customize the dataset?

In the `Dataset Zoo` I see > The field [class] should be 0. Only single-class multi-object tracking is supported in this version. means only support single-class, **Can I tracking person...

Hello, nice work by the way. Can this work be trained for multi-class-multi-object tracking such as tracking multiple cars, multiple persons and multiple dogs at the same time?

when I run demo.py,it shows ModuleNotFoundError: No module named 'cython_bbox' I have tried pip install cython_bbox,but : ERROR: Command errored out with exit status 1: command: 'C:\ProgramData\Anaconda3\python.exe' -u -c 'import...

![image](https://user-images.githubusercontent.com/65213261/102327221-c2d9e100-3fc8-11eb-84c1-c139314e27d3.png) like this. Could anyone tell me? It is OK?