[ICCV'25 Highlight] Derm1M: A Million‑Scale Vision‑Language Dataset Aligned with Clinical Ontology Knowledge for Dermatology

Vision-language foundation model for zero-shot dermatology

[MICCAI‘25 Early Accept] MAKE: Multi-Aspect Knowledge-Enhanced Vision-Language Pretraining for Zero-shot Dermatological Assessment

[MICCAI '26 Early Accept] Quality-Guided Semi-Supervised Learning for Medical Image Segmentation

On-device AI medical triage for community health workers — offline-first PWA with Gemma 4, QR sharing, async telemedicine, AI explainability, 200+ tests

Skin disease classification using deep learning and ANN-based medical image analysis

Superfast, privacy centric RAG VLM dermatologist.

AI Voice Assistant for dermatology clinic that can manage callings, bookings. Tech Stack: TypeSript, React.js , N8N Automation, Tailwind CSS, Eleven Labs

Fairness Auditing in Dermoscopic AI: Quantified a 55% FNR disparity based on Anatomical Localization (Spurious Bias Audit). Focus on Disentangled Representation learning for Equitable AI.

[ECCV'26] Towards Trustworthy Dermatology MLLMs: A Benchmark and Multimodal Evaluator for Diagnostic Narratives

Open research workspace for federated multimodal alopecia prognosis with latent diffusion, rigorous evaluation, and governance-by-design.

Derma MedGemma — Dermatology Vision-Language Fine-Tuning on Indian Diseases

DermaCare_AI is a research-backed dermatology intelligence and educational platform. Built on evidence-based dermatology literature and clinical guidance, it provides skin condition information, skincare assessments, treatment references, emergency recognition tools, and printable reports with privacy-first local processing.

An ML pipeline to classify skin lesions as malignat or benign with a data dashboard and a API interface tested on the HAM10000

A project to understand and build basic knowledge on dermatologic classification/segmentation