Josh Hamilton

Josh is a PhD candidate in the CompuMAINE Lab studying Biomedical Engineering, expected to graduate in August 2026. He grew up in the town of Alton, Maine, and derives his passion for learning from personal experience with local economic disparities and the loss of loved ones to cancer.

His research develops quantitative and deep learning tools for cancer image analysis across mammography and digital pathology. A central thread is the use of wavelet and fractal methods such as the 2D Wavelet Transform Modulus Maxima (WTMM) multifractal approach to characterize how the tumor microenvironment is organized across multiple size scales. In mammography, this work stratifies dense breast tissue into “active” and “passive” subtypes for improved breast cancer risk prediction, contributing to a patent and a $440,000 NCI R-15 award. In histopathology, he quantifies both tissue and collagen fiber organization in H&E and Second Harmonic Generation imaging of breast and pancreatic cancer.

More recently, Josh’s work has expanded into deep learning for digital pathology, including context-guided segmentation of histopathology images and differentiable superpixel segmentation. He is the lead developer of py2DWTMM, an open-source Python toolkit that re-implements the lab’s core analysis pipeline from legacy C/TCL code. His goal is to deepen our understanding of the tumor microenvironment’s morphology through quantitative tools, with the hope of advancing cancer detection, risk prediction, and ultimately patient care.

In his free time, Josh enjoys playing drum kit with his music-major friends and competing in esports such as Super Smash Bros. Melee.

Selected Publications

  • Juybari J, Hamilton J, Das SS, Chen C, Khalil A, Zhu Y. Differentiable Laplacian Matrix Guided Superpixel Segmentation. CVPR 2026 — Oral Presentation (top 3.3% of accepted papers), in press.
  • Juybari J, Hamilton J, Chen C, et al. Context-guided segmentation for histopathologic cancer segmentation. Scientific Reports 15, 5404 (2025). Featured by the NCI.
  • Hamilton J, Breggia A, Fitzgerald TL, Jones MA, Brooks PC, Tilbury K, Khalil A. Multiscale anisotropy analysis of second-harmonic generation collagen imaging of human pancreatic cancer. Frontiers in Oncology 12:991850 (2022).
  • Miner J, Emmerling C, Hamilton J, et al. Targeting the D93 cryptic collagen epitope alters integrin α2β1-dependent cellular migration and collagen remodeling in metastatic breast cancer. Scientific Reports (2026), in press.

LinkedIn Profile: https://www.linkedin.com/in/josh-david-hamilton/