MPViT: Multi-Path Vision Transformer for Dense Prediction

• URL: http://arxiv.org/abs/2112.11010v1
• Date: Tue, 21 Dec 2021 06:34:50 GMT
• Title: MPViT: Multi-Path Vision Transformer for Dense Prediction
• Authors: Youngwan Lee, Jonghee Kim, Jeff Willette, Sung Ju Hwang
• Abstract summary: Vision Transformers (ViTs) build a simple multi-stage structure for multi-scale representation with single-scale patches. OuriTs scaling from tiny(5M) to base(73M) consistently achieve superior performance over state-of-the-art Vision Transformers.
• Score: 43.89623453679854
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Dense computer vision tasks such as object detection and segmentation require effective multi-scale feature representation for detecting or classifying objects or regions with varying sizes. While Convolutional Neural Networks (CNNs) have been the dominant architectures for such tasks, recently introduced Vision Transformers (ViTs) aim to replace them as a backbone. Similar to CNNs, ViTs build a simple multi-stage structure (i.e., fine-to-coarse) for multi-scale representation with single-scale patches. In this work, with a different perspective from existing Transformers, we explore multi-scale patch embedding and multi-path structure, constructing the Multi-Path Vision Transformer (MPViT). MPViT embeds features of the same size~(i.e., sequence length) with patches of different scales simultaneously by using overlapping convolutional patch embedding. Tokens of different scales are then independently fed into the Transformer encoders via multiple paths and the resulting features are aggregated, enabling both fine and coarse feature representations at the same feature level. Thanks to the diverse, multi-scale feature representations, our MPViTs scaling from tiny~(5M) to base~(73M) consistently achieve superior performance over state-of-the-art Vision Transformers on ImageNet classification, object detection, instance segmentation, and semantic segmentation. These extensive results demonstrate that MPViT can serve as a versatile backbone network for various vision tasks. Code will be made publicly available at \url{https://git.io/MPViT}.

Related papers

• M2Former: Multi-Scale Patch Selection for Fine-Grained Visual Recognition [4.621578854541836]
  We propose multi-scale patch selection (MSPS) to improve the multi-scale capabilities of existing ViT-based models. Specifically, MSPS selects salient patches of different scales at different stages of a vision Transformer (MS-ViT) In addition, we introduce class token transfer (CTT) and multi-scale cross-attention (MSCA) to model cross-scale interactions between selected multi-scale patches and fully reflect them in model decisions.
  arXiv  Detail & Related papers  (2023-08-04T06:41:35Z)
• Vision Transformer with Convolutions Architecture Search [72.70461709267497]
  We propose an architecture search method-Vision Transformer with Convolutions Architecture Search (VTCAS) The high-performance backbone network searched by VTCAS introduces the desirable features of convolutional neural networks into the Transformer architecture. It enhances the robustness of the neural network for object recognition, especially in the low illumination indoor scene.
  arXiv  Detail & Related papers  (2022-03-20T02:59:51Z)
• SimViT: Exploring a Simple Vision Transformer with sliding windows [3.3107339588116123]
  We introduce a vision Transformer named SimViT, to incorporate spatial structure and local information into the vision Transformers. SimViT extracts multi-scale hierarchical features from different layers for dense prediction tasks. Our SimViT-Micro only needs 3.3M parameters to achieve 71.1% top-1 accuracy on ImageNet-1k dataset.
  arXiv  Detail & Related papers  (2021-12-24T15:18:20Z)
• Vision Transformer with Progressive Sampling [73.60630716500154]
  We propose an iterative and progressive sampling strategy to locate discriminative regions. When trained from scratch on ImageNet, PS-ViT performs 3.8% higher than the vanilla ViT in terms of top-1 accuracy.
  arXiv  Detail & Related papers  (2021-08-03T18:04:31Z)
• ViTAE: Vision Transformer Advanced by Exploring Intrinsic Inductive Bias [76.16156833138038]
  We propose a novel Vision Transformer Advanced by Exploring intrinsic IB from convolutions, ie, ViTAE. ViTAE has several spatial pyramid reduction modules to downsample and embed the input image into tokens with rich multi-scale context. In each transformer layer, ViTAE has a convolution block in parallel to the multi-head self-attention module, whose features are fused and fed into the feed-forward network.
  arXiv  Detail & Related papers  (2021-06-07T05:31:06Z)
• Less is More: Pay Less Attention in Vision Transformers [61.05787583247392]
  Less attention vIsion Transformer builds upon the fact that convolutions, fully-connected layers, and self-attentions have almost equivalent mathematical expressions for processing image patch sequences. The proposed LIT achieves promising performance on image recognition tasks, including image classification, object detection and instance segmentation.
  arXiv  Detail & Related papers  (2021-05-29T05:26:07Z)
• Point Cloud Learning with Transformer [2.3204178451683264]
  We introduce a novel framework, called Multi-level Multi-scale Point Transformer (MLMSPT) Specifically, a point pyramid transformer is investigated to model features with diverse resolutions or scales. A multi-level transformer module is designed to aggregate contextual information from different levels of each scale and enhance their interactions.
  arXiv  Detail & Related papers  (2021-04-28T08:39:21Z)
• Transformers in Vision: A Survey [101.07348618962111]
  Transformers enable modeling long dependencies between input sequence elements and support parallel processing of sequence. Transformers require minimal inductive biases for their design and are naturally suited as set-functions. This survey aims to provide a comprehensive overview of the Transformer models in the computer vision discipline.
  arXiv  Detail & Related papers  (2021-01-04T18:57:24Z)

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.