how to test own dataset by multi-template?

[oliver8459]

i want to test my own dataset by multi-template, but i found only single template can be select in './demo/main/video/sot_video.py'.
'python3 ./main/test.py --config 'experiments/siamfcpp/test/vot/siamfcpp_alexnet-multi_temp.yaml' ' can use multi-temp but only for open-dataset.

[lzx1413]

You can also use the experiments/siamfcpp/test/vot/siamfcpp_alexnet-multi_temp.yaml in the sot_video.py. Just change the config file path to adopt the multi_temp pipeline.

[oliver8459]

I saw 'self._state['features'] = [self._state['features']] *self._hyper_params['mem_len']' in 'videoanalyst/pipeline/tracker_impl/siamfcpp_track_multi_temp.py', multi_temp only use singe temp to multiple men_len directly, and these templates did not update anymore. So these templates are the same, why did this?

[lzx1413]

The update process with the new feature is here, you can uncomment it to enable this function. https://github.com/MegviiDetection/video_analyst/blob/f4d1bccb1c698961fed3cb70808f1177fab13bdd/videoanalyst/pipeline/tracker_impl/siamfcpp_track_multi_temp.py#L154
We will update the code to fix this error

[oliver8459]

thanks a lot~ another question, if i want to track multi objects, can i put multi object templates into a template pool, then do correlation operation synchronizely in a minibatch.

[lzx1413]

Yes, this is the way while training the network.

[JIANG-CX]

thanks a lot~ another question, if i want to track multi objects, can i put multi object templates into a template pool, then do correlation operation synchronizely in a minibatch.

Hello, could you please explain how to track multi objects in details? Thanks a lot.