Azyri

The project highlights several key features:

• Zero-shot segmentation: Utilizes SAM with a zero-prompt approach, eliminating the need for manual annotations or prompts.
• Sex classification: Employs a low-cost deep learning solution to classify sex from hand X-ray images, both original and segmented.
• Visual explanations: Aggregates attention maps to provide visual explanations of model decisions.
• Novel application: Extends the use case to biological anthropology, specifically for sex estimation of prehistoric handprints.

The primary use case is in medical imaging, where the method improves the analysis of hand X-rays for sex classification, which can be valuable in pediatric patients. Additionally, the research introduces a novel application in biological anthropology, enabling sex estimation from prehistoric handprints. This demonstrates the versatility of the approach across different domains.

The project provides access to the research paper, the code on GitHub, and the dataset on Kaggle. These resources allow researchers and practitioners to explore the methodology, reproduce the results, and potentially apply the technique to their own data. The availability of these materials supports further research and collaboration in the field.