Open dongxinyu1030 opened 1 year ago
Hi, if you want to track all cars in the video, we recommend using demo_instseg.ipynb
to get better results. The WebUI now only supports the method that tracking every new objects appearing in the video, rather than tracking objects of interest.
Hi yamy, thank you for responding, I'll try it.
On Sat, May 20, 2023 at 10:25 PM yamy-cheng @.***> wrote:
Hi, if you want to track all cars in the video, we recommend using demo_instseg.ipynb to get better results. The WebUI now only supports the method that tracking every new objects appearing in the video, rather than tracking objects of interest.
— Reply to this email directly, view it on GitHub https://github.com/z-x-yang/Segment-and-Track-Anything/issues/48#issuecomment-1555922989, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANV72NXFQ3CG3SL45Y5G4W3XHDH4PANCNFSM6AAAAAAYIULILE . You are receiving this because you authored the thread.Message ID: @.***>
Hi, if you want to track all cars in the video, we recommend using
demo_instseg.ipynb
to get better results. The WebUI now only supports the method that tracking every new objects appearing in the video, rather than tracking objects of interest.
Hi yamy.If I want to get all the cars that appear in the video, not just the ones that appear in the first frame, which one should I use, the webui effect doesn't look good
Hi, if you want to track all cars in the video, we recommend using
demo_instseg.ipynb
to get better results. The WebUI now only supports the method that tracking every new objects appearing in the video, rather than tracking objects of interest.Hi yamy.If I want to get all the cars that appear in the video, not just the ones that appear in the first frame, which one should I use, the webui effect doesn't look good
According to your described requirements, better results can be obtained by using the demo_instseg.ipynb
.
Hi, I tried to run web UI demo with the cars video, I used text prompt to label all cars in the first frame and expect the tracker could track them correctly in the following frames. However, the tracking result seems not correct. Not sure if it is caused by unsuitable parameter setting. Could anyone help me with this? :) Checkpoints used: sam_vit_b_01ec64, R50_DeAOTL_PRE_YTB_DAV, groundingdino_swint_ogc Here is the result: Here are all parameters I used:
My email: kyle.dongxinyu@gmail.com