mlsanityData to 14 Oct 2026

Stylistic-STORM (ST-STORM) : Perceiving the Semantic Nature of Appearance

arXiv 2604.16086 ↗First posted 17 Apr 2026cs.CV cs.AI cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

One of the dominant paradigms in self-supervised learning (SSL), illustrated by MoCo or DINO, aims to produce robust representations by capturing features that are insensitive to certain image transformations such as illumination, or geometric changes. This strategy is appropriate when the objective is to recognize objects independently of their appearance. However, it becomes counterproductive as soon as appearance itself constitutes the discriminative signal. In weather analysis, for example, rain streaks, snow granularity, atmospheric scattering, as well as reflections and halos, are not noise: they carry the essential information. In critical applications such as autonomous driving, ignoring these cues is risky, since grip and visibility depend directly on ground conditions and atmospheric conditions. We introduce ST-STORM, a hybrid SSL framework that treats appearance (style) as a semantic modality to be disentangled from content. Our architecture explicitly separates two latent streams, regulated by gating mechanisms. The Content branch aims at a stable semantic representation through a JEPA scheme coupled with a contrastive objective, promoting invariance to appearance variations. In parallel, the Style branch is constrained to capture appearance signatures (textures, contrasts, scattering) through feature prediction and reconstruction under an adversarial constraint. We evaluate ST-STORM on several tasks, including object classification (ImageNet-1K), fine-grained weather characterization, and melanoma detection (ISIC 2024 Challenge). The results show that the Style branch effectively isolates complex appearance phenomena (F1=97% on Multi-Weather and F1=94% on ISIC 2024 with 10% labeled data), without degrading the semantic performance (F1=80% on ImageNet-1K) of the Content branch, and improves the preservation of critical appearance

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 28 of its 45 references that are in the map. First posted on arXiv 17 Apr 2026.

Builds on

28 of 45 references.

References · 28

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,364
  3. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  4. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,513
  5. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
  6. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  7. 2022Generative Adversarial Networkscited by 5,627
  8. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,750
  9. 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,527
  10. 2021BEiT: BERT Pre-Training of Image Transformerscited by 2,375
  11. 2021An Empirical Study of Training Self-Supervised Vision Transformerscited by 1,616
  12. 2025DINOv3cited by 1,414
16 more
  1. 2020Big Self-Supervised Models are Strong Semi-Supervised Learnerscited by 1,380
  2. 2019Semantic Image Synthesis with Spatially-Adaptive Normalizationcited by 1,299
  3. 2023Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecturecited by 1,039
  4. 2025V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningcited by 809
  5. 2020Contrastive Learning for Unpaired Image-to-Image Translationcited by 639
  6. 2020What Should Not Be Contrastive in Contrastive Learningcited by 226
  7. 2023A-JEPA: Joint-Embedding Predictive Architecture Can Listencited by 53
  8. 2016Deep TEN: Texture Encoding Networkcited by 44
  9. 2025Joint Embedding vs Reconstruction: Provable Benefits of Latent Space Prediction for Self Supervised Learningcited by 36
  10. 2022Stare at What You See: Masked Image Modeling without Reconstructioncited by 30
  11. 2023PixMIM: Rethinking Pixel Reconstruction in Masked Image Modelingcited by 28
  12. 2025In Pursuit of Pixel Supervision for Visual Pre-trainingcited by 13
  13. 2023Amortised Invariance Learning for Contrastive Self-Supervisioncited by 8
  14. 2024Understanding the Role of Equivariance in Self-supervised Learningcited by 7
  15. 2023ALADIN-NST: Self-supervised disentangled representation learning of artistic style through Neural Style Transfercited by 3
  16. 2024Real-Time Weather Image Classification with SVMcited by 2

Built on by

Not cited yet.