MAM-CLIP: Vision-Language Pretraining on Mammography Atlases for BI-RADS Classification

• URL: http://arxiv.org/abs/2605.19359v1
• Date: Tue, 19 May 2026 04:42:05 GMT
• Title: MAM-CLIP: Vision-Language Pretraining on Mammography Atlases for BI-RADS Classification
• Abstract summary: Deep learning methods have demonstrated promising results in predicting BI-RADS scores from mammography images.<n>We curated 2313 mammogram images and their corresponding captions from two mammography atlases.<n>We fine-tune the vision encoder on two datasets for BI-RADS prediction, achieving superior performance compared with models trained without this pretraining.
• Score: 2.7579377082303673
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Deep learning methods have demonstrated promising results in predicting BI-RADS scores from mammography images. However, the interpretation of these images can vary, leading to discrepancies even among radiologists. Given the inherent complexity of mammograms, training classification models solely on image labels often yields limited performance. To address this challenge, we curated 2313 mammogram images and their corresponding captions from two mammography atlases. Our proposed approach employs a multi-modal model that uses a pretrained PubMedBERT as the language component. By training this model on image-text pairs with contrastive learning, we enable the vision encoder to absorb the rich information contained in the captions, thereby improving its understanding of mammography findings. We then fine-tune the vision encoder on two datasets for BI-RADS prediction, achieving superior performance compared with models trained without this pretraining, particularly when labeled samples are scarce. The improvement in the 3-class average F1 score ranges from +1% to +14%: a +1% increase with 40K training samples, and a +14% increase with 1K samples. Furthermore, our experiments reveal that 2K image-text pairs from mammography atlases can be more informative than 2K labeled samples for label prediction, with an average margin of +1.1% when more than 10K training samples are available. Overall, our work provides a vision-language model for mammography and highlights the value of textual information from mammography atlases. In addition, we publicly release preprocessed mammography images of the TEKNOFEST dataset. The training code, pre-trained model weights, data extraction scripts, and the released dataset are publicly available at: https://github.com/igulluk/MAM-CLIP

Related papers

• Should VLMs be Pre-trained with Image Data? [54.50406730361859]
  We find that pre-training with a mixture of image and text data allows models to perform better on vision-language tasks.<n>On an average of 6 diverse tasks, we find that for a 1B model, introducing visual tokens 80% of the way through pre-training results in a 2% average improvement over introducing visual tokens to a fully pre-trained model.
  arXiv  Detail & Related papers  (2025-03-10T17:58:19Z)
• MedFILIP: Medical Fine-grained Language-Image Pre-training [11.894318326422054]
  Existing methods struggle to accurately characterize associations between images and diseases.<n>MedFILIP introduces medical image-specific knowledge through contrastive learning.<n>For single-label, multi-label, and fine-grained classification, our model achieves state-of-the-art performance.
  arXiv  Detail & Related papers  (2025-01-18T14:08:33Z)
• ExGra-Med: Extended Context Graph Alignment for Medical Vision-Language Models [95.47808515575382]
  ExGra-Med is a novel framework for vision-language integration in medical AI.<n>It aligns images, instruction responses, and extended captions in the latent space, advancing semantic grounding and cross-modal coherence.<n>It matches LLaVA-Med's performance using just 10% of the pre-training data, achieving a 20.13% gain on VQA-RAD and approaching full-data performance.
  arXiv  Detail & Related papers  (2024-10-03T15:52:03Z)
• CT-GLIP: 3D Grounded Language-Image Pretraining with CT Scans and Radiology Reports for Full-Body Scenarios [53.94122089629544]
  We introduce CT-GLIP (Grounded Language-Image Pretraining with CT scans), a novel method that constructs organ-level image-text pairs to enhance multimodal contrastive learning. Our method, trained on a multimodal CT dataset comprising 44,011 organ-level vision-text pairs from 17,702 patients across 104 organs, demonstrates it can identify organs and abnormalities in a zero-shot manner using natural languages.
  arXiv  Detail & Related papers  (2024-04-23T17:59:01Z)
• CXR-CLIP: Toward Large Scale Chest X-ray Language-Image Pre-training [6.292642131180376]
  In this paper, we tackle the lack of image-text data in chest X-ray by expanding image-label pair as image-text pair via general prompt. We also design two contrastive losses, named ICL and TCL, for learning study-level characteristics of medical images and reports. Our model outperforms the state-of-the-art models trained under the same conditions.
  arXiv  Detail & Related papers  (2023-10-20T05:44:55Z)
• Unified Medical Image-Text-Label Contrastive Learning With Continuous Prompt [3.218449686637963]
  We propose a unified Image-Text-Label contrastive learning framework based on continuous prompts. We demonstrate through sufficient experiments that the Unified Medical Contrastive Learning framework exhibits excellent performance on several downstream tasks.
  arXiv  Detail & Related papers  (2023-07-12T05:19:10Z)
• MOFI: Learning Image Representations from Noisy Entity Annotated Images [47.6984817573981]
  We present MOFI, a new vision foundation model designed to learn image representations from noisy entity annotated images. We introduce a new approach to automatically assign entity labels to images from noisy image-text pairs. Our approach involves employing a named entity recognition model to extract entities from the alt-text, and then using a CLIP model to select the correct entities as labels of the paired image.
  arXiv  Detail & Related papers  (2023-06-13T17:51:18Z)
• Vision-Language Modelling For Radiological Imaging and Reports In The Low Data Regime [70.04389979779195]
  This paper explores training medical vision-language models (VLMs) where the visual and language inputs are embedded into a common space. We explore several candidate methods to improve low-data performance, including adapting generic pre-trained models to novel image and text domains. Using text-to-image retrieval as a benchmark, we evaluate the performance of these methods with variable sized training datasets of paired chest X-rays and radiological reports.
  arXiv  Detail & Related papers  (2023-03-30T18:20:00Z)
• Significantly improving zero-shot X-ray pathology classification via fine-tuning pre-trained image-text encoders [50.689585476660554]
  We propose a new fine-tuning strategy that includes positive-pair loss relaxation and random sentence sampling. Our approach consistently improves overall zero-shot pathology classification across four chest X-ray datasets and three pre-trained models.
  arXiv  Detail & Related papers  (2022-12-14T06:04:18Z)
• Self-Supervised Learning as a Means To Reduce the Need for Labeled Data in Medical Image Analysis [64.4093648042484]
  We use a dataset of chest X-ray images with bounding box labels for 13 different classes of anomalies. We show that it is possible to achieve similar performance to a fully supervised model in terms of mean average precision and accuracy with only 60% of the labeled data.
  arXiv  Detail & Related papers  (2022-06-01T09:20:30Z)
• mc-BEiT: Multi-choice Discretization for Image BERT Pre-training [52.04866462439979]
  Image BERT pre-training with masked image modeling (MIM) is a popular practice to cope with self-supervised representation learning. We introduce an improved BERT-style image pre-training method, namely mc-BEiT, which performs MIM proxy tasks towards eased and refined multi-choice training objectives.
  arXiv  Detail & Related papers  (2022-03-29T09:08:18Z)
• Weakly-supervised Generative Adversarial Networks for medical image classification [1.479639149658596]
  We propose a novel medical image classification algorithm called Weakly-Supervised Generative Adversarial Networks (WSGAN) WSGAN only uses a small number of real images without labels to generate fake images or mask images to enlarge the sample size of the training set. We show that WSGAN can obtain relatively high learning performance by using few labeled and unlabeled data.
  arXiv  Detail & Related papers  (2021-11-29T15:38:48Z)
• Self-supervised Image-text Pre-training With Mixed Data In Chest X-rays [10.398175542736285]
  We introduce an image-text pre-training framework that can learn from mixed data inputs. We demonstrate the feasibility of pre-training across mixed data inputs. We also illustrate the benefits of adopting such pre-trained models in 3 chest X-ray applications.
  arXiv  Detail & Related papers  (2021-03-30T01:48:46Z)

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.