search

shengyuhao / DIVOTrack

A Novel Dataset and Baseline Method for Cross-View Multi-Object Tracking in DIVerse Open Scenes (IJCV 2024)
Other
82 stars 14 forks source link

Change detector #6

Open DuyTa506 opened 1 year ago

DuyTa506 commented 1 year ago

In MTMC pipline , can we use different backbone for detection such as Yolo or SSD Net ?

Hank0626 commented 1 year ago

Currently, we do not support other detector other than CenterNet.

DuyTa506 commented 1 year ago

I have one more question . in CrossMOT pipline , i want to visualize the image , but when i use the display on class Update , the frame_display function return None when read image by cv2 . I tried to print out the self.seq[view]["image_filenames"][frame_idx - self.seq[view]["min_frame_idx"]] but it return just a "S" character , i thought it had to a path of image ? image image

shengyuhao commented 1 year ago

I have one more question . in CrossMOT pipline , i want to visualize the image , but when i use the display on class Update , the frame_display function return None when read image by cv2 . I tried to print out the self.seq[view]["image_filenames"][frame_idx - self.seq[view]["min_frame_idx"]] but it return just a "S" character , i thought it had to a path of image ? image image

We do not provide the code for visualization right now. Visualization by .txt files after inference is recommended and convenient.