In the next Machine Learning Group seminar, we have a talk by Sepehr Jalali who is a post-doctoral researcher at City.
Date & Time: Wednesday 18th June, 16:00-17:00
Location: Room AG05 (College Building)
Title: Inspirations from human visual cortex for image classification
Abstract: Image classification is one of the most challenging problems in computer science due to different illuminations, viewpoints, occlusions, scale and shift transforms in the images. Hence no computer vision approach has been capable of dealing with all these issues to provide a complete solution. On the other hand, the human visual system is considered a superior model for various visual recognition tasks. Hence it has motivated researchers to study its mechanisms in order to simulate them. In this talk, an overview on a hierarchical model, HMAX (Riesenhuber and Poggio, Nature, 1999) for image classification that is biologically inspired and simulates some known characteristics of visual cores is presented. It is followed by our extensions to it, a new colour model (CQ-HMAX) and a comparison on the classification accuracies and performances on different benchmark datasets.
Bio: Dr Sepehr Jalali received his Bachelor’s degree in computer engineering, software, from the university of Isfahan, Iran, in 2006. He was a research assistant with Isfahan University of Technology for two years. He was then awarded the Singapore International Graduate Award (SINGA) scholarship to pursue his PhD studies in Singapore and worked with “Data Storage Institute (DSI)”, “Institute for Infocomm Research (I2R)” and “Image and Pervasive Access Lab (IPAL)” and received his PhD degree from National University of Singapore (NUS) in the department of Electrical and Computer Engineering (ECE) under topic of “Simulating Hierarchical Structure of Human Visual Cortex for Image Classification” in 2013. He completed a Post-doctoral Research Fellowship with the Tropical Marine Science Institute (TMSI) in National University of Singapore upon completion of his PhD Studies. He is currently a Post-Doctoral Research Assistant with the department of Psychology, City University London, working on a BBSRC funded “Sports Decision Making” project. His research interests include image classification, neural networks, support vector machines, biologically inspired models for image classification, scene classification, action recognition and decision making.