Trained RefineNet as well as custom networks for the semantic segmentation of high-resolution aerial imagery. Primarily used: Matlab, Python, and Keras.

Designed and trained novel neural network architectures for one-shot detection of rare objects in satellite imagery.

Experimented with U-Nets and Mask-RCNN for the segmentation of biomedical images. Work was done using: Python, Tensorflow, Keras, and OpenCV.