Unaiza Ahsan - Georgia Institute of Technology

About me

I am a 2nd year PhD student at Georgia Institute of Technology. I work at the Computational Perception Lab (CPL) at the School of Interactive Computing under the supervision of Prof. Irfan Essa. I completed my Master's from NED University of Engineering & Technology, Karachi, Pakistan in Computer Systems and my thesis was on Human tracking. You can find my detailed CV here.

Research Interests

Social Multimedia analysis, Machine Learning based Social Computing, Computer Vision applications on Social Platforms

Latest News

Research Project

Social Event Clustering using Kernel Canonical Correlation

Social Event Clustering using Kernel Canonical Correlation Analysis

Sharing user experiences in the form of photographs, tweets, text, audio and/or video has become commonplace in social media. Browsing through uploaded content of a particular event remains cumbersome. It requires a user to initiate textual search query and manually go through a list of resulting images to find relevant information. We propose an automatic clustering algorithm, which given a large collection of images, groups them into a cluster of different events using the image features and the related metadata. We formulate this problem as a kernel canonical correlation clustering problem in which data samples from different modalities or 'views' are projected to a space where correlations between the samples' projections are maximized.

Publications

Oral

  • U. Ahsan, S. Abdul Sattar, H. Noor and M. Zafar, "Multi-cue object detection and tracking for security in complex environments", Proc. SPIE 8391, Automatic Target Recognition XXII (May 1, 2012) [pdf]

Poster

  • U. Ahsan and I. Essa, "Clustering Social Event Images using Kernel Canonical Correlation Analysis", Georgia Tech Research and Innovation Conference (GTRIC 2014) [pdf]

Workshop

  • U. Ahsan, I. Essa, "Clustering Social Event Images using Kernel Canonical Correlation Analysis", Computer Vision and Pattern Recognition Workshop (CVPRW) on Web-scale Vision and Social Media (VSM) 2014 (accepted).

Master's thesis

  • U. Ahsan, S. Abdul Sattar, H. Noor, "Multi-cue Human Detection and Tracking", Master's thesis, NED University of Engineering & Technology, 2011

Course Projects

Improved Object Recognition using Shape Masks

Improved Object Recognition using Shape Masks

The aim of this project was to implement an improved version of object detection using shape masks to exploit spatial relationship between image features to filter out background clutter. This was an implementation of Marszalek and Schmid's IJCV paper "Accurate Object Recognition with Shape Masks".

You can find our project paper here.

Using Hough Forests for Object Detection

Using Hough Forests for Object Detection

The aim of this project was to detect objects by training a Hough forest and then assigning test image patches to particular leaf nodes, and aggregating the probabilities of the patch belonging to the leaf nodes for the final detection result.

You can find our project paper here.

Using Hough Forests for Object Detection

Scene Completion Using Semantically Coherent Regions with Large Image Database

The aim of this project was to perform scene completion using user-specified object classes (to fill the holes). This project was inspired by Hays et al.'s paper "Scene completion using millions of photographs".

You can find our project paper here.

Courses

Fall 2013

Machine Learning (CS 7641): Course Instructor: Le Song.

Spring 2013

Numerical Linear Algebra (CS 6643): Course Instructor: Edmond Chow

Fall 2012

Computer Vision (CS 7495): Course Instructor: Frank Dellaert

Contact

Email:

uahsan3 [at] gatech.edu