Protean Image Exploration and Learning

GRAND Seminar Monday, Feb 26, 2pm, Room: ENGR 3507

Yekaterina Kharitonova
Visiting Assistant Professor
Computer Science Department
Harvey Mudd College

Host:

Vivian G Motti
vmotti@gmu.edu

Abstract:

Visual data is being produced at an unprecedented rate generating more information that we are able to effectively access and process. The accelerated pace of data creation along with an increased intersection of different disciplines creates a need to recognize and adapt solutions from one field to the similar problems in another. My research goal is to create these solutions by implementing robust, general algorithms, which we can release as open-source tools.

Computer vision aims to teach computers to understand the world through image content. As a computer vision researcher, I focus on approaches to detecting, classifying, and accessing information from visual data. In this talk, I will provide examples from my research that show how we can use computer vision techniques, such as homography and RANSAC, for image matching and video compression, and how cutting-edge approaches involving convolutional neural networks (CNNs) can help us analyze deep-sea videos. As part of my future plans, I will discuss how using the algorithms for summarizing deep-sea videos can help us with matching presentation slides to video, and how a processing pipeline for detecting and classifying marine species can be applied to studying plant distribution and bee pollination.

Short Bio:

Dr. Yekaterina Kharitonova is a Visiting Assistant Professor in the Computer Science Department at Harvey Mudd College (Claremont, CA). Her research is in Computer Vision and Machine Learning. Specifically, her work focuses on multimedia processing and understanding, image correspondence, and effective image alignment through fitting geometric models. She leads the PIXL group, which explores interdisciplinary projects that currently involve large-scale image mapping, and object detection and classification.