7 datasets found

The Extended Yale B database consists of 2414 frontal-face images of 38 subjects. Each subject has around 64 images. The images are cropped and normalized to 192 × 168 under...

The Super-CLEVR-3D dataset contains questions explicitly querying 3D understanding including object parts, 3D poses, and occlusions.

The YCB-V dataset contains 21 objects with complex shapes and occlusions.

Re-identification (Re-ID) problems aim to determine corresponding targets across multiple cameras or different locations in the same camera from the gallery set with a given probe.

YouTube-VOS 2018 and 2019 datasets are benchmarks for video object segmentation, featuring various challenging scenarios.

DAVIS 2016 and 2017 datasets are benchmarks for video object segmentation, featuring various challenging scenarios.

MOSE 2023 dataset is a novel VOS benchmark featuring extreme scenarios of the video sequence which are not handled good enough by existing VOS methods.