Is Yolact anchor based or anchor free ?

[Boatsure]

Thanks for the great engineering application research ！

I see the Yolo-seg series models completely borrow the idea of coefficientized masks from Yolact.

In your paper 3.2 section saying about yolov8: The updated Head module embraces a decoupled structure, separating classification and detection heads, and shifts from Anchor-Based to Anchor-Free. And I'm wondering, does the threshold crop step in Yolact makes the yolo-seg architecture back to an Anchor-Based method?

[YinglongDu]

Thank you for your attention, yolov8 uses the method of anchor-free, which is mainly applied to detect branches, and there is no effect on segment branches, indeed Yolov8-seg uses anchor-free in the threshold crop step, I hope it can solve your doubts.:blush:

[an-yongqi]

We would close this issue for now to keep the issue tracker organized. However, if the problem persists or if you have any further questions, please feel free to comment here or open a new issue. We value your input and are happy to assist further.

[Boatsure]

@YinglongDu @an-yongqi Hello again! My question is about the yolact part so I don't open a new issue. How to output the prototype activation maps in the mask branch, fig 2 in your paper? In the yolact project code, it can be done by --mask_proto_debug and function display_lincomb like https://github.com/dbolya/yolact/issues/278#issuecomment-577967607 says. Thank you!