LightAVSeg: Lightweight Audio-Visual Segmentation
For practitioners needing efficient audio-visual segmentation, this work provides a resource-friendly solution that maintains competitive accuracy.
LightAVSeg introduces a lightweight audio-visual segmentation framework that replaces dense cross-modal attention with a decoupled design, achieving 50.4 mIoU on MS3 with 20.5M parameters (1/7 of AVSegFormer) and enabling mobile inference.
Audio-Visual Segmentation (AVS) targets pixel level localization of sounding emitting objects in videos. However, existing models rely on dense cross-modal attention with quadratic computational cost, limiting their suitability for resource efficient deployment. Most efficiency oriented methods focus on backbone reduction and overlook the interaction module as the primary bottleneck. This paper proposes LightAVSeg, a lightweight framework that replaces heavy attention with a decoupled design for semantic filtering and spatial grounding, resulting in interaction costs that scale linearly with spatial resolution. Furthermore, we introduce an auxiliary alignment loss to enforce semantic consistency during training with zero inference overhead. Extensive experiments demonstrate that LightAVSeg achieves a new state-of-the-art among lightweight methods: with 20.5M parameters ~1/7 of AVSegFormer), it reaches 50.4 mIoU on the MS3 benchmark and enables efficient inference on a mobile processor.