Our team secured an NHMRC Ideas Grant with the University of Queensland on "Tracking melanoma at the pixel level: deep image analysis guided by spatial molecular profiling". Dr. Zhen Yu was appointed as CIB, working alongside CIA Prof H. Peter Soyer (UQ) and collaborators from Medical University of Vienna and Dartmouth College. This 4-year project combines 3D total body photography, AI, and genomics to identify melanoma-prone skin regions for early detection.

Our team is partnering with Cancer Council Victoria and Alfred Health's Victorian Melanoma Service on the My SunSmart project, building on the successful ACEMID pilot study (53 participants, 93% app adoption, 89% using UV forecasts). We are developing an AI-enhanced personalisation system that automatically generates tailored sun protection messages based on individual risk profiles, behaviours, and demographics.

DermoGPT is a morphology-grounded dermatological reasoning multimodal large language model with fully open weights and open data. It introduces a novel DermoBench benchmark and outperforms 16 baselines across 11 clinical tasks, including diagnosis, morphology recognition, and clinical reasoning. The model and data are publicly available to support reproducible research in dermatology AI.

Our multimodal vision foundation model for clinical dermatology — pretrained on over 2 million skin images across 4 imaging modalities and evaluated on 28 benchmarks. Reader studies show +11% diagnostic accuracy for dermatologists, +16.5% for non-specialists, and +10.2% improvement in early melanoma detection. Widely covered by Monash University, Mayo Clinic Platform, Healthcare IT News, and international media.

Our team secured an NHMRC 2024 Clinical Trials and Cohort Studies Grant for "AUTOMATEd risk assessment and screening for MElanoma and skin cancer (AUTOMATE-ME)". Led by CIA Assoc Prof Victoria Mar (Alfred Health/Monash), this 5-year cohort study leverages the ACEMID network across 16 health services with over 6,400 participants. Assoc Prof Zongyuan Ge contributes as CIF and will lead AI algorithm development for automated melanoma screening.

New collaboration with Monash Suzhou Research Institute and CAS Institute of Nutrition and Health on automated facial pigmentation analysis. The project recruits 500 subjects across 7 pigmentation types, combining multimodal vision-language models for AI-driven detection, classification, and severity grading with genome-wide association studies to uncover the genetic basis of pigmentation.

Our team collaborates with Professor Lei Zhang from Monash's School of Translational Medicine and Melbourne Sexual Health Centre on a $1.6M NHMRC-funded project to develop AI-assisted diagnostic and prognostic tools for high-resolution anoscopy (HRA) in anal cancer detection. Our research fellow Jiajun Sun will be responsible for delivering the AI solutions. The project partners with University of Melbourne, St Vincent's Hospital, Mater Health, and Western Health, leveraging Monash's MAVERIC supercomputer for training deep learning models on large-scale cancer image datasets.