CMT-DeepLab: Clustering Mask Transformers for Panoptic Segmentation
This addresses the problem of accurate object and stuff segmentation in computer vision, with a novel method that significantly advances performance.
The paper tackled panoptic segmentation by proposing CMT-DeepLab, a transformer-based framework that uses clustering to group pixels, achieving a new state-of-the-art of 55.7% PQ on COCO test-dev with a 4.4% improvement.
We propose Clustering Mask Transformer (CMT-DeepLab), a transformer-based framework for panoptic segmentation designed around clustering. It rethinks the existing transformer architectures used in segmentation and detection; CMT-DeepLab considers the object queries as cluster centers, which fill the role of grouping the pixels when applied to segmentation. The clustering is computed with an alternating procedure, by first assigning pixels to the clusters by their feature affinity, and then updating the cluster centers and pixel features. Together, these operations comprise the Clustering Mask Transformer (CMT) layer, which produces cross-attention that is denser and more consistent with the final segmentation task. CMT-DeepLab improves the performance over prior art significantly by 4.4% PQ, achieving a new state-of-the-art of 55.7% PQ on the COCO test-dev set.