The Densely Annotation Video Segmentation dataset (DAVIS) is a high quality and high resolution densely annotated video segmentation dataset under two resolutions, 480p and 1080p. There are 50 video sequences with 3455 densely annotated frames in pixel level. 30 videos with 2079 frames are for training and 20 videos with 1376 frames are for validation. Source: TENet: Triple Excitation Network for Video Salient Object Detection

Youtube-VOS is a Video Object Segmentation dataset that contains 4,453 videos - 3,471 for training, 474 for validation, and 508 for testing. The training and validation videos have pixel-level ground truth annotations for every 5th frame (6 fps). It also contains Instance Segmentation annotations. It has more than 7,800 unique objects, 190k high-quality manual annotations and more than 340 minutes in duration. Source: CapsuleVOS: Semi-Supervised Video Object Segmentation Using Capsule Routing Image Source: https://youtube-vos.org/

The How2Sign is a multimodal and multiview continuous American Sign Language (ASL) dataset consisting of a parallel corpus of more than 80 hours of sign language videos and a set of corresponding modalities including speech, English transcripts, and depth. A three-hour subset was further recorded in the Panoptic studio enabling detailed 3D pose estimation.

EgoHOS is a labeled dataset consisting of 11243 egocentric images with per-pixel segmentation labels of hands and objects being interacted with during a diverse array of daily activities. The data are collected form multiple sources: 7,458 frames from Ego4D, 2,212 frames from EPIC-KITCHEN, 806 frames from THU-READ, and 350 frames of our own collected egocentric videos with people playing Escape Room. This dataset is designed for tasks including hand state classification, video activity recognition, 3D mesh reconstruction of hand-object interactions, and video inpainting of hand-object foregrounds in egocentric videos. Source: Fine-grained Egocentric Hand-object Segmentation: Dataset, Model, and Applications Image Source: https://arxiv.org/pdf/2208.03826v1.pdf

Are current 3D object tracking methods truely robust enough for low-fidelity depth sensors like the iPhone LiDAR? We introduce DTTD-Mobile (fully compatible w/ YCB toolbox), a new benchmark built on real-world data captured from mobile devices; 18 objects observed in 100 videos with 47,668 sampled frames and 114,143 object annotations. We evaluate several popular methods—including BundleSDF, ES6D, MegaPose, and DenseFusion—and highlight their limitations in this challenging setting. Keywords: ObjectPoseEstimation, MobileAI, EdgeAI, ARVR, CVPR

The Free-Form Video Inpainting dataset is a dataset used for training and evaluation video inpainting models. It consists of 1940 videos from the YouTube-VOS dataset and 12,600 videos from the YouTube-BoundingBoxes. Source: https://arxiv.org/abs/1904.10247 Image Source: https://github.com/amjltc295/Free-Form-Video-Inpainting

KITTI360-EX is a dataset for outer- and inner FoV expansion. It contains 76k pinhole images as well as 76k spherical images and is used for beyond-FoV estimation. Source: FlowLens: Seeing Beyond the FoV via Flow-guided Clip-Recurrent Transformer Image Source: https://arxiv.org/pdf/2211.11293v1.pdf

Diagnostic Evaluation of Video Inpainting on Landscapes (DEVIL) benchmark is composed of a curated video/occlusion mask dataset and a comprehensive evaluation scheme

QST contains 1,167 video clips that are cut out from 216 time-lapse 4K videos collected from YouTube, which can be used for a variety of tasks, such as (high-resolution) video generation, (high-resolution) video prediction, (high-resolution) image generation, texture generation, image inpainting, image/video super-resolution, image/video colorization, image/video animating, etc. Each short clip contains multiple frames (from a minimum of 58 frames to a maximum of 1,200 frames, a total of 285,446 frames), and the resolution of each frame is more than 1,024 x 1,024. Specifically, QST consists of a training set (containing 1000 clips, totally 244,930 frames), a validation set (containing 100 clips, totally 23,200 frames), and a testing set (containing 67 clips, totally 17,316 frames). Click here (Key: qst1) to download the QST dataset.

The Inpainting dataset consists of synchronized Labeled image and LiDAR scanned point clouds. It's captured by HESAI Pandora All-in-One Sensing Kit. It is collected under various lighting conditions and traffic densities in Beijing, China.

We provide video sequences with annotated object masks for video inpainting. The resolution is 3840 x 2160.

The benchmark for VPData, the largest video inpainting dataset, which comprises over 390K clips (> 866.7 hours) and features precise masks and detailed video captions.

The largest video inpainting dataset comprises over 390K clips (> 866.7 hours), featuring precise masks and detailed video captions.

Dataset for the DREAMING - Diminished Reality for Emerging Applications in Medicine through Inpainting Challenge! The DREAMING Challenge is part of ISBI 2024. Participate for Awards, Money Prices, Proceedings Paper, Top-Tier Challenge Summary Paper! More information about the challenge and registration can be found on the challenge platform! The DREAMING training dataset contains 100 diverse, photorealistic surgery scenes with unique patients, surroundings and obstructions from dynamically moving and static surgical instruments and hands holding them. We also provide the masks defining the area to be inpainted for each image.

G2LP Wire-removal Dataset in G2LP-Net: Global to Local Progressive Video Inpainting Network # Wire-removal Dataset The WRV dataset has been specifically curated for the challenges of video inpainting in irregularly slender regions. It encompasses 150 video clips (now 190) that are extracted from movies and TV series. What sets the WRV dataset apart is its real-world scenarios where video frames may not have fixed foregrounds, or might even showcase multiple dynamic foregrounds such as action sequences involving several individuals. A distinct feature of this dataset is the naturally occurring masks due to dynamic thin wires and small auxiliary props commonly used in martial arts stunts. This makes the dataset particularly representative of genuine challenges found in film post-production. The primary challenge with the WRV dataset lies in effectively utilizing contextual information. Due to the irregular and slender nature of the to-be-inpainted regions, leveraging surrounding contextual details is paramount to achieving authentic inpainting results. ## Wire-removal Dataset Link (480P version) Baidu Disk Google Drive ## Wire-removal Dataset Link(4K version) Waiting for updates ...... ## Citation If this research benefits your work or involves the use of this dataset, please consider citing the following paper: @article{ji2022g2lp, title={G2LP-Net: Global to Local Progressive Video Inpainting Network}, author={Ji, Zhong and Hou, Jiacheng and Su, Yimu and Pang, Yanwei and Li, Xuelong}, journal={IEEE Transactions on Circuits and Systems for Video Technology}, volume={33}, number={3}, pages={1082--1092}, year={2022}, publisher={IEEE} } ## License This project is licensed under the MIT License. Please refer to the LICENSE file for detailed terms.