Main Menu (Mobile)- Block

Main Menu - Block

Gary Huang

people_blocks-people_blocks_name_and_labs | block
People / Gary HuangScientific Computing Software
people_blocks-people_blocks_action_picture | block
people_blocks-people_blocks_bio_summary | block
Senior Software Engineer
people_blocks-people_blocks_biography | block

Biography

My research interests are in computer vision, machine learning, and deep learning.  I did my graduate work at the Vision Lab at the University of Massachusetts - Amherst, under the supervision of Professor Erik Learned-Miller.  The main focus of my work was developing weakly-supervised learning methods for improving unconstrained face verification.  More information can be found at my UMass-Amherst webpage, as well as Labeled Faces in the Wild, a database for studying unconstrained face verification.

After graduating in 2012, I came to Janelia Research Campus and joined the Viren Jain Lab, working on developing algorithms for machine segmentation and reconstruction from electron microscopy (EM) images.  In particular, we developed an architecture called Deep and Wide Multiscale Recursive Networks (DAWMR) for segmentation of EM volumes.

In 2014, I joined the FlyEM project team at Janelia, where I extended DAWMR for the purposes of automated synapse prediction, evaluated a pipeline for fully-automated synapse prediction on a large EM data set, and am currently working on deep learning methods such as convolutional neural networks to improve the accuracy and throughput of automated methods for understanding EM data.

node:field_education | entity_field

Education

PhD, Computer science, University of Massachusetts Amherst
BS; MS, Stanford University
janelia7_blocks-janelia7_featured_blocks | block

Featured Publications

Gary B. Huang, Viren Jain. Deep and Wide Multiscale Recursive Networks for Robust Image Labeling. International Conference on Learning Representations, 2014.
John A. Bogovic, Gary B. Huang, Viren Jain. Learned versus Hand-Designed Feature Representations for 3d Agglomeration. International Conference on Learning Representations, 2014.