arrow
返回

Unsupervised Modality-Transferable Video Highlight Detection With Representation Activation Sequence Learning

delete2024-01-01
delete0
delete
OA
AI
T
Tingtian Li *
Z
Zixun Sun
X
Xinyu Xiao
DOI:10.1109/TIP.2024.3372469delete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
Identifying highlight moments of raw video materials is crucial for improving the efficiency of editing videos that are pervasive on internet platforms. However, the extensive work of manually labeling footage has created obstacles to applying supervised methods to videos of unseen categories. The absence of an audio modality that contains valuable cues for highlight detection in many videos also makes it difficult to use multimodal strategies. In this paper, we propose a novel model with cross-modal perception for unsupervised highlight detection. The proposed model learns representations with visual-audio level semantics from image-audio pair data via a self-reconstruction task. To achieve unsupervised highlight detection, we investigate the latent representations of the network and propose the representation activation sequence learning (RASL) module with k-point contrastive learning to learn significant representation activations. To connect the visual modality with the audio modality, we use the symmetric contrastive learning (SCL) module to learn the paired visual and audio representations. Furthermore, an auxiliary task of masked feature vector sequence (FVS) reconstruction is simultaneously conducted during pretraining for representation enhancement. During inference, the cross-modal pretrained model can generate representations with paired visual-audio semantics given only the visual modality. The RASL module is used to output the highlight scores. The experimental results show that the proposed framework achieves superior performance compared to other state-of-the-art approaches.
Keyword:
Unsupervised
highlight detection
multimodal
representation activation sequence

期刊

IEEE Transactions on Image Processing 封面图
IEEE Transactions on Image Processing
IF:
13.7
论文数:
1.0W
被引数:
8.4W

机构

T
Tencent
学者数:
1.1K
论文数: 898
被引数: 5
引用论文

引用论文

Solution structure of the Lewis x oligosaccharide determined by NMR spectroscopy and molecular dynamics simulations
err2002-05-01
err0
PREAI
errKristine E. Miller; Chaitali Mukhopadhyay; Perseveranda Cagas; C. Allen Bush
err分享
err收藏
Structural, electronic, optical and elastic properties of XLa2S4 (X = Ba; Ca): Ab initio study
err2019-04-01
err0
PREAI
errM. Batouche; T. Seddik; Tuan V. Vu; Khang D. Pham; Hien D. Tong; S. Uğur; G. Uğur; A.A. Lavrentyev; O.Y. Khyzhun; Hai L. Luong; A. Belfedal
err分享
err收藏
Novel de novo variant in EBF3 is likely to impact DNA binding in a patient with a neurodevelopmental disorder and expanded phenotypes: patient report, in silico functional assessment, and review of published cases
err2017-03-10
err0
errOAAI
errPatrick R. Blackburn; Sarah S. Barnett; Michael T. Zimmermann; Margot A. Cousin; Charu Kaiwar; Filippo Pinto e Vairo; Zhiyv Niu; Matthew J. Ferber; Raul A. Urrutia; Duygu Selcen; Eric W. Klee; Pavel N. Pichurin
err分享
err收藏
学者 查看更多内容