Dual-Level Cross-Modal Contrastive Clustering

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Haixin, Li, Yongjun, Huang, Dong
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917780122501120
author Zhang, Haixin
Li, Yongjun
Huang, Dong
author_facet Zhang, Haixin
Li, Yongjun
Huang, Dong
contents Image clustering, which involves grouping images into different clusters without labels, is a key task in unsupervised learning. Although previous deep clustering methods have achieved remarkable results, they only explore the intrinsic information of the image itself but overlook external supervision knowledge to improve the semantic understanding of images. Recently, visual-language pre-trained model on large-scale datasets have been used in various downstream tasks and have achieved great results. However, there is a gap between visual representation learning and textual semantic learning, and how to properly utilize the representation of two different modalities for clustering is still a big challenge. To tackle the challenges, we propose a novel image clustering framwork, named Dual-level Cross-Modal Contrastive Clustering (DXMC). Firstly, external textual information is introduced for constructing a semantic space which is adopted to generate image-text pairs. Secondly, the image-text pairs are respectively sent to pre-trained image and text encoder to obtain image and text embeddings which subsquently are fed into four well-designed networks. Thirdly, dual-level cross-modal contrastive learning is conducted between discriminative representations of different modalities and distinct level. Extensive experimental results on five benchmark datasets demonstrate the superiority of our proposed method.
format Preprint
id arxiv_https___arxiv_org_abs_2409_04561
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Dual-Level Cross-Modal Contrastive Clustering
Zhang, Haixin
Li, Yongjun
Huang, Dong
Computer Vision and Pattern Recognition
Image clustering, which involves grouping images into different clusters without labels, is a key task in unsupervised learning. Although previous deep clustering methods have achieved remarkable results, they only explore the intrinsic information of the image itself but overlook external supervision knowledge to improve the semantic understanding of images. Recently, visual-language pre-trained model on large-scale datasets have been used in various downstream tasks and have achieved great results. However, there is a gap between visual representation learning and textual semantic learning, and how to properly utilize the representation of two different modalities for clustering is still a big challenge. To tackle the challenges, we propose a novel image clustering framwork, named Dual-level Cross-Modal Contrastive Clustering (DXMC). Firstly, external textual information is introduced for constructing a semantic space which is adopted to generate image-text pairs. Secondly, the image-text pairs are respectively sent to pre-trained image and text encoder to obtain image and text embeddings which subsquently are fed into four well-designed networks. Thirdly, dual-level cross-modal contrastive learning is conducted between discriminative representations of different modalities and distinct level. Extensive experimental results on five benchmark datasets demonstrate the superiority of our proposed method.
title Dual-Level Cross-Modal Contrastive Clustering
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2409.04561