Snarcy
Abstract
RedDino, a self-supervised foundation model using DINOv2, excels in RBC shape classification and generalization, addressing challenges in computational hematology.
Red blood cells (RBCs) are essential to human health, and their precise morphological analysis is important for diagnosing hematological disorders. Despite the promise of foundation models in medical diagnostics, comprehensive AI solutions for RBC analysis remain scarce. We present RedDino, a self-supervised foundation model designed for RBC image analysis. RedDino uses an RBC-specific adaptation of the DINOv2 self-supervised learning framework and is trained on a curated dataset of 1.25 million RBC images from diverse acquisition modalities and sources. Extensive evaluations show that RedDino outperforms existing state-of-the-art models on RBC shape classification. Through assessments including linear probing and nearest neighbor classification, we confirm its strong feature representations and generalization ability. Our main contributions are: (1) a foundation model tailored for RBC analysis, (2) ablation studies exploring DINOv2 configurations for RBC modeling, and (3) a detailed evaluation of generalization performance. RedDino addresses key challenges in computational hematology by capturing nuanced morphological features, advancing the development of reliable diagnostic tools. The source code and pretrained models for RedDino are available at https://github.com/Snarci/RedDino, and the pretrained models can be downloaded from our Hugging Face collection at https://huggingface.co/collections/Snarcy/reddino-689a13e29241d2e5690202fc
Community
We introduce RedDino, our self-supervised foundation model designed specifically for red blood cell (RBC) analysis. Building on a customized DINOv2 backbone, we trained on the largest RBC dataset to date over 1.25 million RBC images and more than 3 million segmented cells from 18 datasets, spanning multiple imaging modalities and staining protocols.
Across multiple datasets (Elsafty, Chula, DSE), RedDino outperforms ResNet50, DinoBloom, and vanilla DINOv2 by +2–3% in weighted F1, balanced accuracy, and accuracy, while showing strong generalization to out-of-distribution data.
With RedDino, we set a new benchmark for automated RBC morphological analysis and provide a robust, generalizable backbone for future hematological AI applications.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- Benchmarking Foundation Models for Mitotic Figure Classification (2025)
- ABCD: Automatic Blood Cell Detection via Attention-Guided Improved YOLOX (2025)
- AI-Driven Cytomorphology Image Synthesis for Medical Diagnostics (2025)
- Benchmarking Deep Learning and Vision Foundation Models for Atypical vs. Normal Mitosis Classification with Cross-Dataset Evaluation (2025)
- Transfer Learning with EfficientNet for Accurate Leukemia Cell Classification (2025)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment:
@librarian-bot
recommend
Models citing this paper 3
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper