Shuai Li

I tried WITH_PYTHON_LAYER := 1, but still facing the same error although I successfully build the caffe and pycaffe.

@Gerjo Thank you Gerjo for your nice advice! Really works for me using default gcc and g++ 5.4.0 in Ubuntu16.04. I installed g++ 4.8.0 and it really helps a lot!

@corentin87 Thank you,saved me lots of time!

@YangBain @sjoycode Oh my godness, how foolish I am. I just messed up the root of vatic. In my case. I just modified the 000-configure file in apache. and I...

You can perform tracktor's tracking on raw sequences, and use their output unlabeled detections as filtered detections

I think setting the maximum temporal gap between nodes to a small value, e.g. 10 frames in the original paper should help. Multicut itself is NP-hard, meaning the bigger the...

You can do NMS to filter multiple observations associated with one target. And for the multiple ID Switches, you can use a pretrained Car re-identification appearance model.

The author's output are slightly different from what MOT challenge needs. You can convert it according to MOT's requirement.

[libprotobuf WARNING google/protobuf/io/coded_stream.cc:78] The total number of bytes read was 703909485 F1120 08:52:31.197538 10638 io.cpp:67] Check failed: proto.SerializeToOstream(&output) *** Check failure stack trace: *** Aborted (core dumped) This is my...

There seems to have no restriction on video's resolution. In my case, the tool is not able to extract a video in 640 x 480, but after I convert the...