mlsanityData to 14 Oct 2026

Stable Single-Pixel Contrastive Learning for Semantic and Geometric Tasks

arXiv 2512.04970 ↗First posted 4 Dec 2025cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

We pilot a family of stable contrastive losses for learning pixel-level representations that jointly capture semantic and geometric information. Our approach maps each pixel of an image to an overcomplete descriptor that is both view-invariant and semantically meaningful. It enables precise point-correspondence across images without requiring momentum-based teacher-student training. Two experiments in synthetic 2D and 3D environments demonstrate the properties of our loss and the resulting overcomplete representations.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 27 of its 40 references that are in the map. First posted on arXiv 4 Dec 2025.

Builds on

27 of 40 references.

References · 27

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  4. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  5. 2023Segment Anythingcited by 8,234
  6. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  7. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  8. 2021BEiT: BERT Pre-Training of Image Transformerscited by 2,375
  9. 2019Invariant Risk Minimizationcited by 1,854
  10. 2022Vision Transformer Adapter for Dense Predictionscited by 565
  11. 2020Dense Contrastive Learning for Self-Supervised Visual Pre-Trainingcited by 512
  12. 2023Simple diffusion: End-to-end diffusion for high resolution imagescited by 422
15 more
  1. 2021Deep ViT Features as Dense Visual Descriptorscited by 330
  2. 2020Propagate Yourself: Exploring Pixel-Level Consistency for Unsupervised Visual Representation Learningcited by 297
  3. 2016Universal Correspondence Networkcited by 82
  4. 2023Semantic Segmentation using Vision Transformers: A surveycited by 73
  5. 2022A Survey on Label-efficient Deep Image Segmentation: Bridging the Gap between Weak Supervision and Dense Predictioncited by 46
  6. 2024Local Feature Matching Using Deep Learning: A Surveycited by 26
  7. 2022Location-Aware Self-Supervised Transformers for Semantic Segmentationcited by 20
  8. 2025Visual question answering: from early developments to recent advances -- a surveycited by 19
  9. 2024Near, far: Patch-ordering enhances vision foundation models' scene understandingcited by 15
  10. 2020Exploring Overcomplete Representations for Single Image Deraining using CNNscited by 11
  11. 2023From Saliency to DINO: Saliency-guided Vision Transformer for Few-shot Keypoint Detectioncited by 8
  12. 2020Overcomplete Deep Subspace Clustering Networkscited by 7
  13. 2022A Critical Analysis of Image-based Camera Pose Estimation Techniquescited by 7
  14. 2022Learning Dense Object Descriptors from Multiple Views for Low-shot Category Generalizationcited by 6
  15. 2024Aligning Motion-Blurred Images Using Contrastive Learning on Overcomplete Pixelscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Beyond Objective Expressivity: Geometry Preservation in Multimodal Contrastive Learning
  2. 2026Pixel-Perfect Visual Geometry Estimation
  3. 2025PointDico: Contrastive 3D Representation Learning Guided by Diffusion Models
  4. 2026Expressivity In Multimodal Contrastive Learning
  5. 2026Continual Learning with Vision-Language Models via Semantic-Geometry Preservation
  6. 2026Learning Representations to Guide Pixel Diffusion Transformers
  7. 2022Unifying Vision-Language Representation Space with Single-tower Transformer
  8. 2022Perceptual Grouping in Contrastive Vision-Language Models