Consulting

Pixel Scientia Labs I provide consulting services through my company, Pixel Scientia Labs, automating analysis of scientific image data sets. We apply cutting-edge research in image recognition and machine learning to develop custom solutions to unique and challenging image-based problems. We have extensive experience with interdisciplinary R&D projects and are focused on applications for scientific discovery such as digital pathology and planetary science.

Research

Selected Publications

(2019). Deep Multi-View Learning via Task-Optimal CCA. arXiv preprint 1907.07739.

PDF

(2019). Discriminative Representations for Heterogeneous Images and Multimodal Data. Doctoral Dissertation.

PDF Slides

(2018). Image analysis with deep learning to predict breast cancer grade, ER status, histologic subtype, and intrinsic subtype. npj Breast Cancer.

PDF Code

(2018). Multiple Instance Learning for Heterogeneous Images: Training a CNN for Histopathology. Proc. MICCAI.

PDF Poster

(2010). Scene classification of images and video via semantic segmentation. Proc. CVPR Workshop on Perceptual Organization in Computer Vision.

PDF Slides

(2007). Multi-scale Features for Detection and Segmentation of Rocks in Mars Images. Proc. CVPR.

PDF Poster

(2006). Automatic Rock Detection and Classification in Natural Scenes. Masters Thesis.

PDF Slides

Software

These methods were developed for classifying larger images with intra-image heterogeneity. The class of each image is not dictated by some small to-be-identified region but by the overall appearance - perhaps the average class of smaller regions or some more complex relationship.

Coming soon.

Coming soon.

Writing

A Deeper Understanding of Breast Cancer
Jun. 27, 2019
AI finds new insights into molecular tumor properties using images of cells and tissue
The same technology that powers Siri and face recognition on your iPhone has also found success in medicine. By automatically analyzing microscopic images of breast tumor biopsies, artificial intelligence may one day help guide cancer treatments.

Press

Scientists train a computer to classify breast cancer tumors
Nov. 19, 2018
In a study published in the journal NPJ Breast Cancer, researchers reported they used a form of artificial intelligence called machine learning, or deep learning, to train a computer to identify certain features of breast cancer tumors from images. The computer also identified the tumor type based on complex molecular and genomic features, which a pathologist can’t yet identify from a picture alone. They believe this approach, while still in its early stages, could eventually lead to cost savings for the clinic and in breast cancer research.