Insights, discoveries, and thoughts on Computer Vision, Medical Image Analysis, and AI
Exploring how multi-task learning architectures can significantly improve performance on medical image segmentation and classification tasks, particularly when dealing with imbalanced datasets.
Read MoreA technical overview of our work on variational graph autoencoders with residual connections and their applications in social network analysis.
Read MoreHow we improved cell tracking efficiency using novel autoencoder architectures for time-series analysis in label-free optical microscopy.
Read MorePractical insights into model compression and quantization techniques that can accelerate inference from 2 fps to 18 fps on surveillance systems.
Read MoreDeep dive into our layer clustering approach for identifying the most informative hint points in knowledge distillation networks.
Read MoreReflections on transitioning from industry research roles at Vodafone and Arcelik to pursuing a PhD in Computer Science at UMBC.
Read MoreTry a different keyword or clear the filters to see all posts.