Zoya Shafique

I am a second year Ph.D. student in Electrical Engineering in the Media Lab at The City College of New York. My research focus is in computer vision. I am especially interested in studying data-centric deep learning approaches for computer vision applications and in developing efficient and robust feature spaces from data to combat the data-hungry nature of present day architectures.

Email  /  LinkedIn  /  Google Scholar  /  Github  /  CV        

• [2023.05.12] Our paper on nonverbal cue recognition in videos is accepted at the Women in Computer Vision workshop at CVPR 2023 as an oral presentation!

My main research focus is in developing efficient methods to extract robust features from data in order to combat the data-hungry nature of modern day deep learning architectures.

We introduce a dataset for nonverbal communication cue recognition in videos and develop a baseline for facial expression recognition. Our baseline achieves comparable performance as previous state of the art methods on the Aff-Wild2 dataset.

Below are various projects I have worked on as part of my graduate courses.

We present a self-supervised LSTM to classify the underlying morals (as based on Moral Foundational Theory) in different Reddit posts.

We present a thorough analysis of the trends and patterns in the Yelp Dataset and build both collaborative and content based filters to recommend restaurants.

We employ both traditional computer vision methods and a deep neural network for localizing doors in indoor scenes.

We use k-means clustering to create color maps from RGB images and apply those color maps to the corresponding regions in grayscale images.

Teaching Assistant, Digital Image Processing, Spring 2023

This website is cloned from Jon Barron's repository,