AI-Powered Medical Annotation Platform
Developed a scalable web-based interface to streamline medical imaging workflows. Leveraged the Segment Anything Model (SAM) to enable high-precision, semi-automated segmentation of complex medical data, and designed a modular backend architecture allowing users to upload and evaluate custom deep learning models for real-time inference.
JavaScriptTypeScriptSAMPython
Automated Hematology Analysis System
Architected a multi-class deep learning pipeline utilizing state-of-the-art architectures (ResNet, DenseNet121) for the classification of White Blood Cells (WBC) and Red Blood Cells (RBC). Developed computer vision algorithms for automated platelet (thrombocyte) detection, precise cell counting, and automated medical reporting logic.
Deep LearningComputer VisionPythonPyTorch