RadJEPA: Radiology Encoder for Chest X-Rays via Joint Embedding Predictive Architecture

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Khan, Anas Anwarul Haq, Husain, Mariam, Jalan, Pratik, Jadhav, Kshitij
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913162161291264
author Khan, Anas Anwarul Haq
Husain, Mariam
Jalan, Pratik
Jadhav, Kshitij
author_facet Khan, Anas Anwarul Haq
Husain, Mariam
Jalan, Pratik
Jadhav, Kshitij
contents Vision-language pretraining has driven much of the recent progress in medical image representation learning, but this paradigm is constrained by the availability of paired image-text data and by the reporting bias of clinical narratives. We ask whether competitive radiology encoders can be learned without any language supervision. We introduce RadJEPA, a self-supervised framework built on a Joint Embedding Predictive Architecture and pretrained on approximately 840K unlabeled chest X-ray images. The model learns to predict latent representations of masked target regions from a visible context region, an objective that differs from both image-text contrastive pretraining and DINO-style self-distillation by explicitly modelling conditional structure in representation space. We evaluate RadJEPA primarily on radiology report generation with a frozen Vicuna-7B decoder, and additionally substitute its encoder into four widely used vision-language backbones (MedLLaVA, Qwen-2.5, BLIP-2, and Phi-4). For completeness we also report disease classification and semantic segmentation results. Across two datasets and four metrics, RadJEPA matches or exceeds the strongest image-only and vision-language baselines while using a ViT-B/14 backbone at 224 x 224 resolution.
format Preprint
id arxiv_https___arxiv_org_abs_2601_15891
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle RadJEPA: Radiology Encoder for Chest X-Rays via Joint Embedding Predictive Architecture
Khan, Anas Anwarul Haq
Husain, Mariam
Jalan, Pratik
Jadhav, Kshitij
Computer Vision and Pattern Recognition
Vision-language pretraining has driven much of the recent progress in medical image representation learning, but this paradigm is constrained by the availability of paired image-text data and by the reporting bias of clinical narratives. We ask whether competitive radiology encoders can be learned without any language supervision. We introduce RadJEPA, a self-supervised framework built on a Joint Embedding Predictive Architecture and pretrained on approximately 840K unlabeled chest X-ray images. The model learns to predict latent representations of masked target regions from a visible context region, an objective that differs from both image-text contrastive pretraining and DINO-style self-distillation by explicitly modelling conditional structure in representation space. We evaluate RadJEPA primarily on radiology report generation with a frozen Vicuna-7B decoder, and additionally substitute its encoder into four widely used vision-language backbones (MedLLaVA, Qwen-2.5, BLIP-2, and Phi-4). For completeness we also report disease classification and semantic segmentation results. Across two datasets and four metrics, RadJEPA matches or exceeds the strongest image-only and vision-language baselines while using a ViT-B/14 backbone at 224 x 224 resolution.
title RadJEPA: Radiology Encoder for Chest X-Rays via Joint Embedding Predictive Architecture
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.15891