Unsupervised Keypoints from Pretrained Diffusion Models

• URL: http://arxiv.org/abs/2312.00065v3
• Date: Tue, 21 May 2024 22:37:11 GMT
• Title: Unsupervised Keypoints from Pretrained Diffusion Models
• Authors: Eric Hedlin, Gopal Sharma, Shweta Mahajan, Xingzhe He, Hossam Isack, Abhishek Kar Helge Rhodin, Andrea Tagliasacchi, Kwang Moo Yi,
• Abstract summary: We leverage the emergent knowledge within text-to-image diffusion models, towards more robust unsupervised keypoints. Our core idea is to find text embeddings that would cause the generative model to consistently attend to compact regions in images. We validate our performance on multiple datasets: the CelebA, CUB-200-2011, Tai-Chi-HD, DeepFashion, and Human3.6m datasets.
• Score: 31.147785019795347
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Unsupervised learning of keypoints and landmarks has seen significant progress with the help of modern neural network architectures, but performance is yet to match the supervised counterpart, making their practicability questionable. We leverage the emergent knowledge within text-to-image diffusion models, towards more robust unsupervised keypoints. Our core idea is to find text embeddings that would cause the generative model to consistently attend to compact regions in images (i.e. keypoints). To do so, we simply optimize the text embedding such that the cross-attention maps within the denoising network are localized as Gaussians with small standard deviations. We validate our performance on multiple datasets: the CelebA, CUB-200-2011, Tai-Chi-HD, DeepFashion, and Human3.6m datasets. We achieve significantly improved accuracy, sometimes even outperforming supervised ones, particularly for data that is non-aligned and less curated. Our code is publicly available and can be found through our project page: https://ubc-vision.github.io/StableKeypoints/

Related papers

• Diffusion-based Data Augmentation for Object Counting Problems [62.63346162144445]
  We develop a pipeline that utilizes a diffusion model to generate extensive training data. We are the first to generate images conditioned on a location dot map with a diffusion model. Our proposed counting loss for the diffusion model effectively minimizes the discrepancies between the location dot map and the crowd images generated.
  arXiv  Detail & Related papers  (2024-01-25T07:28:22Z)
• VoxelKP: A Voxel-based Network Architecture for Human Keypoint Estimation in LiDAR Data [53.638818890966036]
  textitVoxelKP is a novel fully sparse network architecture tailored for human keypoint estimation in LiDAR data. We introduce sparse box-attention to focus on learning spatial correlations between keypoints within each human instance. We incorporate a spatial encoding to leverage absolute 3D coordinates when projecting 3D voxels to a 2D grid encoding a bird's eye view.
  arXiv  Detail & Related papers  (2023-12-11T23:50:14Z)
• Learning Feature Matching via Matchable Keypoint-Assisted Graph Neural Network [52.29330138835208]
  Accurately matching local features between a pair of images is a challenging computer vision task. Previous studies typically use attention based graph neural networks (GNNs) with fully-connected graphs over keypoints within/across images. We propose MaKeGNN, a sparse attention-based GNN architecture which bypasses non-repeatable keypoints and leverages matchable ones to guide message passing.
  arXiv  Detail & Related papers  (2023-07-04T02:50:44Z)
• Deep vanishing point detection: Geometric priors make dataset variations vanish [24.348651041697114]
  Deep learning has improved vanishing point detection in images. Yet, deep networks require expensive annotated datasets trained on costly hardware. Here, we address these issues by injecting deep vanishing point detection networks with prior knowledge.
  arXiv  Detail & Related papers  (2022-03-16T12:34:27Z)
• Keypoint Message Passing for Video-based Person Re-Identification [106.41022426556776]
  Video-based person re-identification (re-ID) is an important technique in visual surveillance systems which aims to match video snippets of people captured by different cameras. Existing methods are mostly based on convolutional neural networks (CNNs), whose building blocks either process local neighbor pixels at a time, or, when 3D convolutions are used to model temporal information, suffer from the misalignment problem caused by person movement. In this paper, we propose to overcome the limitations of normal convolutions with a human-oriented graph method. Specifically, features located at person joint keypoints are extracted and connected as a spatial-temporal graph
  arXiv  Detail & Related papers  (2021-11-16T08:01:16Z)
• BDC: Bounding-Box Deep Calibration for High Performance Face Detection [11.593495085674345]
  Modern CNN-based face detectors have achieved tremendous strides due to large annotated datasets. misaligned results with high detection confidence but low localization accuracy restrict the further improvement of detection performance. We propose a novel Bounding-Box Deep (BDC) method to reasonably replace inconsistent annotations with model predicted bounding-boxes.
  arXiv  Detail & Related papers  (2021-10-08T04:41:41Z)
• Accurate Grid Keypoint Learning for Efficient Video Prediction [87.71109421608232]
  Keypoint-based video prediction methods can consume substantial computing resources in training and deployment. In this paper, we design a new grid keypoint learning framework, aiming at a robust and explainable intermediate keypoint representation for long-term efficient video prediction. Our method outperforms the state-ofthe-art video prediction methods while saves 98% more than computing resources.
  arXiv  Detail & Related papers  (2021-07-28T05:04:30Z)
• SA-Det3D: Self-Attention Based Context-Aware 3D Object Detection [9.924083358178239]
  We propose two variants of self-attention for contextual modeling in 3D object detection. We first incorporate the pairwise self-attention mechanism into the current state-of-the-art BEV, voxel and point-based detectors. Next, we propose a self-attention variant that samples a subset of the most representative features by learning deformations over randomly sampled locations.
  arXiv  Detail & Related papers  (2021-01-07T18:30:32Z)
• Pixel-in-Pixel Net: Towards Efficient Facial Landmark Detection in the Wild [104.61677518999976]
  We propose Pixel-in-Pixel Net (PIPNet) for facial landmark detection. The proposed model is equipped with a novel detection head based on heatmap regression. To further improve the cross-domain generalization capability of PIPNet, we propose self-training with curriculum.
  arXiv  Detail & Related papers  (2020-03-08T12:23:42Z)

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.