What is Point Supervision Worth in Video Instance Segmentation?
This work addresses the annotation cost problem for video instance segmentation researchers and practitioners, offering an incremental improvement in efficiency.
The paper tackles the problem of reducing expensive dense annotations in video instance segmentation by using only one point per object per frame during training, achieving high-quality mask predictions close to fully supervised models with competitive performance on three benchmarks.
Video instance segmentation (VIS) is a challenging vision task that aims to detect, segment, and track objects in videos. Conventional VIS methods rely on densely-annotated object masks which are expensive. We reduce the human annotations to only one point for each object in a video frame during training, and obtain high-quality mask predictions close to fully supervised models. Our proposed training method consists of a class-agnostic proposal generation module to provide rich negative samples and a spatio-temporal point-based matcher to match the object queries with the provided point annotations. Comprehensive experiments on three VIS benchmarks demonstrate competitive performance of the proposed framework, nearly matching fully supervised methods.