Jun Hao Liew

TL;DR: This work proposes a new deep framework, called Regional Interactive Segmentation Network (RIS-Net), to expand the field-of-view of the given inputs to capture the local regional information surrounding them for local refinement.

TL;DR: This work systematically investigates performance drop of the state-of-the-art two-stage instance segmentation model Mask R-CNN on the recent long-tail LVIS dataset, and unveils that a major cause is the inaccurate classification of object proposals.

TL;DR: Wang et al. as mentioned in this paper propose a simple calibration framework to more effectively alleviate classification head bias with a bi-level class balanced sampling approach, which significantly boosts the performance of instance segmentation for tail classes on the recent LVIS dataset and our sampled COCO-LT dataset.

TL;DR: A simple two-stage solution is proposed that enables the Inside-Outside Guidance approach to produce high quality instance segmentation masks from existing datasets with off-the-shelf bounding boxes such as ImageNet and Open Images, demonstrating the superiority of the IOG as an annotation tool.