Lorinc Balog

Papers

Total Citations

H-Index

About

Lorinc Balog is a researcher at the forefront of biologically inspired computer vision, specializing in data-efficient deep learning and egocentric perception. His work bridges neuroscience and artificial intelligence, most notably through the development of a space-variant visual pathway model that mimics the retino-cortical mapping found in mammalian vision. This approach, detailed in his highly cited 2019 paper (8 citations), enables deep convolutional neural networks to process visual data with remarkable efficiency—achieving significant data reduction while preserving scale and rotation invariance. Balog’s foundational 2017 work (6 citations) introduced the concept of a "software retina," a biologically motivated preprocessing layer that dramatically reduces computational load for robot vision and wearable camera systems. His contributions are particularly impactful for autonomous systems operating in resource-constrained environments, where traditional deep learning approaches falter. By demonstrating that biological visual processing principles can enhance artificial perception, Balog has opened new pathways for energy-efficient, real-time computer vision. His research continues to influence the design of next-generation perception systems, making him a key figure in the convergence of computational neuroscience and applied machine learning.