OCC-MLLM-V2: Joint understanding and generation for occluded objects via multi-modal token learning
Authors:
- Chaoyi WangCWChaoyi Wang
- Qingdong HeQHQingdong He
- Yifan YangYYYifan Yang
- Jun PeiJPJun Pei
- Lijie XiaLXLijie Xia
- Jianpo LiuJLJianpo Liu
- Baoqing LiBLBaoqing Li
- Xinhan DiXDXinhan Di
Open Access
- 1 July 2026
- journal article
- Published by Elsevier in Journal of Visual Communication and Image Representation
- Vol. 119
Abstract
No abstract availableFunding Information
- Giant
- Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences (6142804230104)
This publication has 25 references indexed in Scilit:
- A Survey of Embodied Learning for Object-centric Robotic ManipulationMachine Intelligence Research, 2025
- Generative Multimodal Models are In-Context Learners2024
- Unified-IO 2: Scaling Autoregressive Multimodal Models with Vision, Language, Audio, and Action2024
- gSDF: Geometry-Driven Signed Distance Functions for 3D Hand-Object Reconstruction2023
- What's in your hands? 3D Reconstruction of Generic Objects in Hands2022
- DexYCB: A Benchmark for Capturing Hand Grasping of Objects2021
- HOnnotate: A Method for 3D Annotation of Hand and Object Poses2020
- Learning Joint Reconstruction of Hands and Manipulated Objects2019
- Semantic Amodal Segmentation2017
- Capturing Hands in Action Using Discriminative Salient Points and Physics SimulationInternational Journal of Computer Vision, 2016