Neil Patrick Del Gallego is currently the Vice Chair of the Department of Software Technology, College of Computer Studies, at De La Salle University. Before working at DLSU, Neil has been in the PH gaming industry for five years at Anino Games Inc., and Playlab Inc.
Aside from his professional game development experience, his research interests include image processing, computer vision, augmented reality, and data mining. Recently, he published research works in super-resolution, image correction, and image dehazing.
He is among the faculty members of DLSU, College of Computer Studies, actively building the Interactive Entertainment undergraduate program, partnered with Ubisoft Philippines. He is also the lab head of the Graphics, Animation, Multimedia, and Entertainment (GAME) lab at DLSU CCS.
Convolutional Neural Networks for Image Correction
Convolutional neural networks (CNNs) are a class of artificial neural networks that has become a standard approach for many image and video recognition tasks. In this talk, Neil will show a different take on CNNs, where it is applied to an image correction problem. Given a distorted image (e.g. skewed, perspectively distorted, misaligned), the goal is to produce a corrected image. This talk is based on a peer-reviewed research article that I have written. The talk is expected to cover theoretical concepts on the image formation model and our proposed method of using multiple CNNs, trained to rectify/correct a given distorted image using PyTorch and OpenCV as libraries.