Aruni RoyChowdhury

Learn More
We propose bilinear models, a recognition architecture that consists of two feature extractors whose outputs are multiplied using outer product at each location of the image and pooled to obtain an image descriptor. This architecture can model local pairwise feature interactions in a translationally invariant manner which is particularly useful for(More)
The recent explosive growth in convolutional neural network (CNN) research has produced a variety of new archi-tectures for deep learning. One intriguing new architecture is the bilinear CNN (B-CNN), which has shown dramatic performance gains on certain fine-grained recognition problems [13]. We apply this new CNN to the challenging new face recognition(More)
In 2007, Labeled Faces in the Wild was released in an effort to spur research in face recognition, specifically for the problem of face verification with un-constrained images. Since that time, more than 50 papers have been published that improve upon this benchmark in some respect. A remarkably wide variety of innovative methods have been developed to(More)
Overview We study the effect of signed square-root and`2 normal-ization on the accuracy. We also present visualizations on the learned B-CNN models and the most confused classes for the cars and aircraft datasets. Tab. 1 shows the results of the B-CNN (D,M) model w/o fine-tuning using various normalizations of the bilinear vector. These experiments on the(More)
– Research Assistant (Sep. 2014 – current): Face recognition project under IARPA's Janus program. Distinguishing weather phenomena from bird migration patterns in radar imagery.
  • 1