How to generate the class-agnostic segmentation results (0-1 mask) based on the pre-trained model?

mmaaz60 / mvits_for_class_agnostic_od

[ECCV'22] Official repository of paper titled "Class-agnostic Object Detection with Multi-modal Transformer".

MIT License

299 stars 25 forks source link

How to generate the class-agnostic segmentation results (0-1 mask) based on the pre-trained model? #23

Closed cjfcsjt closed 1 year ago

cjfcsjt commented 1 year ago

Thanks for your excellent work. Could you please tell me how to generate the masks of the saliency object given the pre-trained model?

mmaaz60 commented 1 year ago

Hi @cjfcsjt,

Thank You for your interest in our work. MAVL only provides the class-agnostic bounding boxes. For salient object detection, we use the pretrained COCO Mask RCNN head to generate the segmentation masks given the salient object bounding boxes. Please look at the Section D.5 of our paper for more details.

Thanks