KINS

The KINS dataset/benchmark contains images used to evaluate the performance of image composition methods, particularly in scenarios involving imperfectly segmented or incomplete objects.

Papers using KINS (2)

ViTA-Seg: Vision Transformer for Amodal Segmentation in Robotics2025Deep Occlusion-Aware Instance Segmentation with Overlapping BiLayers2021 · 8 cites