Video Instance Shadow Detection Under the Sun and Sky

TL;DR

This paper introduces ViShadow, a semi-supervised framework for video instance shadow detection that combines labeled images and unlabeled videos to improve tracking and shadow-object association.

Contribution

The paper presents a novel semi-supervised approach with a two-stage training pipeline and a new dataset and metric for video instance shadow detection.

Findings

ViShadow improves shadow tracking accuracy in videos.

The framework enhances applications like shadow editing and inpainting.

New dataset SOBA-VID and metric SOAP-VID facilitate evaluation.

Abstract

Instance shadow detection, crucial for applications such as photo editing and light direction estimation, has undergone significant advancements in predicting shadow instances, object instances, and their associations. The extension of this task to videos presents challenges in annotating diverse video data and addressing complexities arising from occlusion and temporary disappearances within associations. In response to these challenges, we introduce ViShadow, a semi-supervised video instance shadow detection framework that leverages both labeled image data and unlabeled video data for training. ViShadow features a two-stage training pipeline: the first stage, utilizing labeled image data, identifies shadow and object instances through contrastive learning for cross-frame pairing. The second stage employs unlabeled videos, incorporating an associated cycle consistency loss to enhance tracking ability. A retrieval mechanism is introduced to manage temporary disappearances, ensuring tracking continuity. The SOBA-VID dataset, comprising unlabeled training videos and labeled testing videos, along with the SOAP-VID metric, is introduced for the quantitative evaluation of VISD solutions. The effectiveness of ViShadow is further demonstrated through various video-level applications such as video inpainting, instance cloning, shadow editing, and text-instructed shadow-object manipulation.