github.com/healthonrails/annolid

github.com/healthonrails/annolid

I created the dataset with one of our lab's videos. I will check and will share the link in the Github repo. It is very easy to label and create a new dataset based on your own video. If you can share your video with me, I will label and create a sample dataset for you in a few hours. Best!

@@annolid drive.google.com/drive/folders/1ePf13lgxG8mcba1EnfhAPtmgdJOhR9Si?usp=sharing the one named Novel Object Test 1.mp4 and I would love to know how to do that as well for future work :D

@@mohamedfadl789 Nice video! I will label some frames and will share the dataset with you. FYI: You can check the following videos for how to create your own dataset. Extract frames: kzbin.info/www/bejne/nKmTo6yXjdWXe8U; Label objects: kzbin.info/www/bejne/sImblIh_ord8ldU; Save Labels: kzbin.info/www/bejne/ppObiWeYprGFkJI; Create COCO dataset: kzbin.info/www/bejne/laCxeJaVaZmYkJo.

@@mohamedfadl789 Here is an example COCO format dataset based on your video drive.google.com/file/d/1YTbyFE6TYImfrxB-1kSFZcVu1PjzAfyj/view?usp=sharing.

Here is the Github repo: github.com/healthonrails/annolid.

@@annolid thanks I wrote the link from the video already :D

@@annolid btw does it work with the novel object test and if it does, does the background have to be colored or can it work on white colored floors ?

@@mohamedfadl789 You need to label some frames from your own video. Then you can train a model base on your dataset. The background can be any color. FYI: here are some videos about the process. healthonrails.github.io/annolid/guide.html#video-guide

Thanks