Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation

被引:0
|
作者
Yan, Shilin [1 ,2 ]
Zhang, Renrui [2 ,3 ]
Guo, Ziyu [3 ]
Chen, Wenchao [1 ]
Zhang, Wei [1 ]
Li, Hongyang [2 ]
Qiao, Yu [2 ]
Dong, Hao [4 ,5 ]
He, Zhongjiang [6 ]
Gao, Peng [2 ]
机构
[1] Fudan Univ, Sch Comp Sci, Shanghai, Peoples R China
[2] Shanghai Artificial Intelligence Lab, Shanghai, Peoples R China
[3] Chinese Univ Hong Kong, Hong Kong, Peoples R China
[4] Peking Univ, Sch CS, Beijing, Peoples R China
[5] PKU agibot Lab, Beijing, Peoples R China
[6] China Telecom Corp Ltd, Data&AI Technol Co, Beijing, Peoples R China
基金
中国国家自然科学基金;
关键词
D O I
暂无
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Recently, video object segmentation (VOS) referred by multi-modal signals, e.g., language and audio, has evoked increasing attention in both industry and academia. It is challenging for exploring the semantic alignment within modalities and the visual correspondence across frames. However, existing methods adopt separate network architectures for different modalities, and neglect the inter-frame temporal interaction with references. In this paper, we propose MUTR, a Multi-modal Unified Temporal transformer for Referring video object segmentation. With a unified framework for the first time, MUTR adopts a DETR-style transformer and is capable of segmenting video objects designated by either text or audio reference. Specifically, we introduce two strategies to fully explore the temporal relations between videos and multi-modal signals. Firstly, for low-level temporal aggregation before the transformer, we enable the multi-modal references to capture multi-scale visual cues from consecutive video frames. This effectively endows the text or audio signals with temporal knowledge and boosts the semantic alignment between modalities. Secondly, for high-level temporal interaction after the transformer, we conduct inter-frame feature communication for different object embeddings, contributing to better object-wise correspondence for tracking along the video. On Ref-YouTube-VOS and AVSBench datasets with respective text and audio references, MUTR achieves +4.2% and +8.7% J&F improvements to state-of-the-art methods, demonstrating our significance for unified multi-modal VOS. Code is released at https://github.com/OpenGVLab/MUTR.
引用
收藏
页码:6449 / 6457
页数:9
相关论文
共 50 条
  • [11] Multi-modality video shot clustering with tensor representation
    Yanan Liu
    Fei Wu
    Multimedia Tools and Applications, 2009, 41 : 93 - 109
  • [12] Multi-modality video shot clustering with tensor representation
    Liu, Yanan
    Wu, Fei
    MULTIMEDIA TOOLS AND APPLICATIONS, 2009, 41 (01) : 93 - 109
  • [13] A Unified Algorithm for Object Tracking and Segmentation and its Application on Intelligent Video Surveillance for Transformer Substation
    Chen X.
    Han Y.
    Yan Y.
    Qi D.
    Shen J.
    Zhongguo Dianji Gongcheng Xuebao/Proceedings of the Chinese Society of Electrical Engineering, 2020, 40 (23): : 7578 - 7586
  • [14] Accounting for Random Regressors: A Unified Approach to Multi-modality Imaging
    Yang, Xue
    Lauzon, Carolyn B.
    Crainiceanu, Ciprian
    Caffo, Brian
    Resnick, Susan M.
    Landman, Bennett A.
    MULTIMODAL BRAIN IMAGE ANALYSIS, 2011, 7012 : 1 - +
  • [15] Homologous point transformer for multi-modality prostate image registration
    Ruchti, Alexander
    Neuwirth, Alexander
    Lowman, Allison K.
    Duenweg, Savannah R.
    LaViolette, Peter S.
    Bukowy, John D.
    PEERJ COMPUTER SCIENCE, 2022, 8
  • [16] Homologous point transformer for multi-modality prostate image registration
    Ruchti, Alexander
    Neuwirth, Alexander
    Lowman, Allison K.
    Duenweg, Savannah R.
    LaViolette, Peter S.
    Bukowy, John D.
    PeerJ Computer Science, 2022, 8
  • [17] TUMOR SEGMENTATION VIA MULTI-MODALITY JOINT DICTIONARY LEARNING
    Wang, Yan
    Yu, Biting
    Wang, Lei
    Zu, Chen
    Luo, Yong
    Wu, Xi
    Yang, Zhipeng
    Zhou, Jiliu
    Zhou, Luping
    2018 IEEE 15TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI 2018), 2018, : 1336 - 1339
  • [18] A Fast Unsupervised Approach for Multi-Modality Surgical Trajectory Segmentation
    Zhao, Hongfa
    Xie, Jiexin
    Shao, Zhenzhou
    Qu, Ying
    Guan, Yong
    Tan, Jindong
    IEEE ACCESS, 2018, 6 : 56411 - 56422
  • [19] Cell Segmenter: A General Framework for Multi-modality Cell Segmentation
    Hu, Kaiwen
    Zhang, Shengxuming
    Jia, Zhijie
    Cheng, Lechao
    Feng, Zunlei
    COMPETITIONS IN NEURAL INFORMATION PROCESSING SYSTEMS, VOL 212, 2022, 212
  • [20] Weakly Supervised Cell Instance Segmentation for Multi-Modality Microscopy
    Xue, Ming
    COMPETITIONS IN NEURAL INFORMATION PROCESSING SYSTEMS, VOL 212, 2022, 212