Cv

Mitigating Mode Collapse by Sidestepping Catastrophic Forgetting

Future Person Localization in First-Person Videos

Developed a novel deep architecture for predicting future locations of people observed in first-person videos. Released a new dataset on human locomotion seen in FPV. Paper under review at CVPR’18.

Compressing U-net using Knowledge Distillation

Designed a knowledge distillation procedure aimed for compressing Fully Convolutional Network (FCN) with skip connections such as U-net and Stacked Hourglass Network. Achieved ~100x model compression.

Emotion Recognition from Static Human facial Images

Implemented various classfiers for human emotion detection in the Emotion Recognition in the Wild Challenge 2016 dataset to compare against results from CNN based approaches. Details on project page.