SPAR: Single-Pass Any-Resolution ViT for Open-vocabulary Segmentation
This addresses computational bottlenecks in dense prediction tasks like segmentation for computer vision applications, representing an incremental improvement in efficiency.
The paper tackles the inefficiency of high-resolution image processing in Vision Transformers for open-vocabulary segmentation by introducing SPAR, a resolution-agnostic feature extractor that improves single-pass baselines by up to 10.5 mIoU and surpasses a sliding-window teacher.
Foundational Vision Transformers (ViTs) have limited effectiveness in tasks requiring fine-grained spatial understanding, due to their fixed pre-training resolution and inherently coarse patch-level representations. These challenges are especially pronounced in dense prediction scenarios, such as open-vocabulary segmentation with ViT-based vision-language models, where high-resolution inputs are essential for accurate pixel-level reasoning. Existing approaches typically process large-resolution images using a sliding-window strategy at the pre-training resolution. While this improves accuracy through finer strides, it comes at a significant computational cost. We introduce SPAR: Single-Pass Any-Resolution ViT, a resolution-agnostic dense feature extractor designed for efficient high-resolution inference. We distill the spatial reasoning capabilities of a finely-strided, sliding-window teacher into a single-pass student using a feature regression loss, without requiring architectural changes or pixel-level supervision. Applied to open-vocabulary segmentation, SPAR improves single-pass baselines by up to 10.5 mIoU and even surpasses the teacher, demonstrating effectiveness in efficient, high-resolution reasoning. Code: https://github.com/naomikombol/SPAR