Siamese Image Modeling for Self-Supervised Vision Representation Learning

• URL: http://arxiv.org/abs/2206.01204v1
• Date: Thu, 2 Jun 2022 17:59:58 GMT
• Title: Siamese Image Modeling for Self-Supervised Vision Representation Learning
• Authors: Chenxin Tao, Xizhou Zhu, Gao Huang, Yu Qiao, Xiaogang Wang, Jifeng Dai
• Abstract summary: Self-supervised learning (SSL) has delivered superior performance on a variety of downstream vision tasks. Two main-stream SSL frameworks have been proposed, i.e., Instance Discrimination (ID) and Masked Image Modeling (MIM) We propose Siamese Image Modeling (SIM), which predicts the dense representations of an augmented view.
• Score: 73.78790119050056
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Self-supervised learning (SSL) has delivered superior performance on a variety of downstream vision tasks. Two main-stream SSL frameworks have been proposed, i.e., Instance Discrimination (ID) and Masked Image Modeling (MIM). ID pulls together the representations of different views from the same image, while avoiding feature collapse. It does well on linear probing but is inferior in detection performance. On the other hand, MIM reconstructs the original content given a masked image. It excels at dense prediction but fails to perform well on linear probing. Their distinctions are caused by neglecting the representation requirements of either semantic alignment or spatial sensitivity. Specifically, we observe that (1) semantic alignment demands semantically similar views to be projected into nearby representation, which can be achieved by contrasting different views with strong augmentations; (2) spatial sensitivity requires to model the local structure within an image. Predicting dense representations with masked image is therefore beneficial because it models the conditional distribution of image content. Driven by these analysis, we propose Siamese Image Modeling (SIM), which predicts the dense representations of an augmented view, based on another masked view from the same image but with different augmentations. Our method uses a Siamese network with two branches. The online branch encodes the first view, and predicts the second view's representation according to the relative positions between these two views. The target branch produces the target by encoding the second view. In this way, we are able to achieve comparable linear probing and dense prediction performances with ID and MIM, respectively. We also demonstrate that decent linear probing result can be obtained without a global loss. Code shall be released.

Related papers

• SA$^2$VP: Spatially Aligned-and-Adapted Visual Prompt [59.280491260635266]
  Methods for visual prompt tuning follow the sequential modeling paradigm stemming from NLP. Mymodel model learns a two-dimensional prompt token map with equal (or scaled) size to the image token map. Our model can conduct individual prompting for different image tokens in a fine-grained manner.
  arXiv  Detail & Related papers  (2023-12-16T08:23:43Z)
• Self-supervised Cross-view Representation Reconstruction for Change Captioning [113.08380679787247]
  Change captioning aims to describe the difference between a pair of similar images. Its key challenge is how to learn a stable difference representation under pseudo changes caused by viewpoint change. We propose a self-supervised cross-view representation reconstruction network.
  arXiv  Detail & Related papers  (2023-09-28T09:28:50Z)
• CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View Completion [20.121597331207276]
  Masked Image Modeling (MIM) has recently been established as a potent pre-training paradigm. In this paper we seek to learn representations that transfer well to a wide variety of 3D vision and lower-level geometric downstream tasks. Our experiments show that our pretext task leads to significantly improved performance for monocular 3D vision downstream tasks.
  arXiv  Detail & Related papers  (2022-10-19T16:50:36Z)
• Dense Semantic Contrast for Self-Supervised Visual Representation Learning [12.636783522731392]
  We present Dense Semantic Contrast (DSC) for modeling semantic category decision boundaries at a dense level. We propose a dense cross-image semantic contrastive learning framework for multi-granularity representation learning. Experimental results show that our DSC model outperforms state-of-the-art methods when transferring to downstream dense prediction tasks.
  arXiv  Detail & Related papers  (2021-09-16T07:04:05Z)
• Seed the Views: Hierarchical Semantic Alignment for Contrastive Representation Learning [116.91819311885166]
  We propose a hierarchical semantic alignment strategy via expanding the views generated by a single image to textbfCross-samples and Multi-level representation. Our method, termed as CsMl, has the ability to integrate multi-level visual representations across samples in a robust way.
  arXiv  Detail & Related papers  (2020-12-04T17:26:24Z)
• Unsupervised Learning of Dense Visual Representations [14.329781842154281]
  We propose View-Agnostic Dense Representation (VADeR) for unsupervised learning of dense representations. VADeR learns pixelwise representations by forcing local features to remain constant over different viewing conditions. Our method outperforms ImageNet supervised pretraining in multiple dense prediction tasks.
  arXiv  Detail & Related papers  (2020-11-11T01:28:11Z)
• Self-Supervised Ranking for Representation Learning [108.38993212650577]
  We present a new framework for self-supervised representation learning by formulating it as a ranking problem in an image retrieval context. We train a representation encoder by maximizing average precision (AP) for ranking, where random views of an image are considered positively related. In principle, by using a ranking criterion, we eliminate reliance on object-centric curated datasets.
  arXiv  Detail & Related papers  (2020-10-14T17:24:56Z)
• Multi-Margin based Decorrelation Learning for Heterogeneous Face Recognition [90.26023388850771]
  This paper presents a deep neural network approach to extract decorrelation representations in a hyperspherical space for cross-domain face images. The proposed framework can be divided into two components: heterogeneous representation network and decorrelation representation learning. Experimental results on two challenging heterogeneous face databases show that our approach achieves superior performance on both verification and recognition tasks.
  arXiv  Detail & Related papers  (2020-05-25T07:01:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.