Ouail Bendidi

@ambr89 I can help you with the translation , I have already translated the 3 new darknet layers to darkflow (need to look for the project cause it's been a...

@qiningonline or you could just make a script that reformat your data to coco format, I believe it's easier that doing step 1 , it's just parsing the values in...

I'm working on an update very soon to get that accuracy up

It's quite simple, this project is just a merge between two different projects : * Darkflow : for object detection (you can use any other object detector for that )...

you just take that bb info , format it into a list of boxes and feed it to deep sort or sort

there are 2 scripts that need to be changed for adding new trackers : * [help.py](https://github.com/bendidi/darkflow/blob/8d164b3c6283efcf3e8ef5c35323e4d3d3850c19/darkflow/net/help.py) : to initialise the tracker and send it to the post-processing unit * [predict.py](https://github.com/bendidi/darkflow/blob/8d164b3c6283efcf3e8ef5c35323e4d3d3850c19/darkflow/net/yolov2/predict.py)...

if you are interested in implementing it feel free to make a pull request, I didn't do it yet

did you check the display FLAG if it's on `True` ?

you could try adding optical flow , with opencv , it works well with occlusions : https://github.com/akshaychawla/Optical-Flow-Tracking----OpenCV i'm working on another version of this project with support for multiple object...

the training scripts for deep sort are not yet released , but you can infer the network architecture from this [script](https://github.com/bendidi/deep_sort/blob/60e082dd132d89d2645f6a4d0c75353a05510176/generate_detections.py) and the training params from their paper , and...