Results 31 to 40 of about 1,683,190 (291)
AdaptFormer: Adapting Vision Transformers for Scalable Visual Recognition [PDF]
Neural Information Processing Systems, 2022 Pretraining Vision Transformers (ViTs) has achieved great success in visual recognition. A following scenario is to adapt a ViT to various image and video recognition tasks.
Shoufa Chen +6 more
semanticscholar +1 more source
Vision Transformers for Single Image Dehazing [PDF]
IEEE Transactions on Image Processing, 2022 Image dehazing is a representative low-level vision task that estimates latent haze-free images from hazy images. In recent years, convolutional neural network-based methods have dominated image dehazing.
Yuda Song, Zhuqing He, Hui Qian, Xin Du
semanticscholar +1 more source
Vision Transformers Need Registers [PDF]
International Conference on Learning Representations, 2023 Transformers have recently emerged as a powerful tool for learning visual representations. In this paper, we identify and characterize artifacts in feature maps of both supervised and self-supervised ViT networks.
Timothée Darcet +3 more
semanticscholar +1 more source
Scaling Vision Transformers to 22 Billion Parameters [PDF]
International Conference on Machine Learning, 2023 The scaling of Transformers has driven breakthrough capabilities for language models. At present, the largest large language models (LLMs) contain upwards of 100B parameters.
Mostafa Dehghani +41 more
semanticscholar +1 more source
Scaling Vision Transformers to Gigapixel Images via Hierarchical Self-Supervised Learning [PDF]
Computer Vision and Pattern Recognition, 2022 Vision Transformers (ViTs) and their multi-scale and hierarchical variations have been successful at capturing image representations but their use has been generally studied for low-resolution images (e.g. 256 × 256, 384 × 384). For gigapixel whole-slide
Richard J. Chen +6 more
semanticscholar +1 more source
EfficientFormer: Vision Transformers at MobileNet Speed [PDF]
Neural Information Processing Systems, 2022 Vision Transformers (ViT) have shown rapid progress in computer vision tasks, achieving promising results on various benchmarks. However, due to the massive number of parameters and model design, \textit{e.g.}, attention mechanism, ViT-based models are ...
Yanyu Li +7 more
semanticscholar +1 more source
ConViT: improving vision transformers with soft convolutional inductive biases [PDF]
International Conference on Machine Learning, 2021 Convolutional architectures have proven to be extremely successful for vision tasks. Their hard inductive biases enable sample-efficient learning, but come at the cost of a potentially lower performance ceiling.
Stéphane d'Ascoli +5 more
semanticscholar +1 more source
Applied Sciences, 2021 Person Re-Identification is an essential task in computer vision, particularly in surveillance applications. The aim is to identify a person based on an input image from surveillance photographs in various scenarios.
Muhammad Tahir, Saeed Anwar
doaj +1 more source
MViTv2: Improved Multiscale Vision Transformers for Classification and Detection [PDF]
Computer Vision and Pattern Recognition, 2021 In this paper, we study Multiscale Vision Transformers (MViTv2) as a unified architecture for image and video classification, as well as object detection.
Yanghao Li +6 more
semanticscholar +1 more source
CMT: Convolutional Neural Networks Meet Vision Transformers [PDF]
Computer Vision and Pattern Recognition, 2021 Vision transformers have been successfully applied to image recognition tasks due to their ability to capture long-range dependencies within an image.
Jianyuan Guo +6 more
semanticscholar +1 more source