Translated data: The field of image segmentation has undergone a transformation driven by deep learning technologies. Mask2Former, a Transformer-based model, excels in semantic, instance, and panoptic segmentation tasks. It boasts superior performance, although it faces FPS limitations on resource-constrained devices. Project link: https://debuggercafe.com/mask2former/