Canonical logit- and feature-based knowledge distillation outperform complex segmentation-specific methods under matched wall-clock compute and achieve near-teacher performance with extended training on Cityscapes and ADE20K.
Masked-attention mask transformer for universal image segmentation
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2representative citing papers
DiTTA distills SAM2 temporal segmentation knowledge into image models via efficient test-time adaptation and a lightweight fusion module to produce annotation-free video semantic segmentation that matches or exceeds fully supervised performance.
citing papers explorer
-
The Surprising Effectiveness of Canonical Knowledge Distillation for Semantic Segmentation
Canonical logit- and feature-based knowledge distillation outperform complex segmentation-specific methods under matched wall-clock compute and achieve near-teacher performance with extended training on Cityscapes and ADE20K.
-
Bootstrapping Video Semantic Segmentation Model via Distillation-assisted Test-Time Adaptation
DiTTA distills SAM2 temporal segmentation knowledge into image models via efficient test-time adaptation and a lightweight fusion module to produce annotation-free video semantic segmentation that matches or exceeds fully supervised performance.