Projects

Iris : Speech to Code

Published:

Speech to code is a hackathon project done at Harvard University during HackHarvard 2018.

Empirical Analysis of Jenkins Pipelines

Published:

This is a course project of CS 540 at UIC. The project involves analysis of large number of Jenkins pipelines and find patterns in those files. A number of research questions are investigated. We were able to find a number of patterns in the pipelines and the detailed report is below. The repository for the project can be found here.

Autograder

Published:

This was a course project for CS 421 : Natural Language Processing at UIC. The goal of this project was to build a automatic grading system for TOEFL essays using NLP techniques and basic NLP building blocks such as POS Tagging, Named Entity Recognition, Spelling correction, grammar error detection and text coherence.

Comprehensive Guide to Gitlab - Jenkins Setup

Published:

The project explores setup of Gitlab and Jenkins and then setting up of DevOps pipeline for all repositories in GitLab in Jenkins automatically using Job DSLs. Further webhooks are automatically added to complete the pipeline. Below is a detailed step by step guide.

MS Apriori

Published:

MS - Apriori is an algorithm to find frequent itemsets from data with multiple minimum support. With the regular Apriori algorithm, frequent itemsets found has only one support regardless of the frequency of the itemsets. This creates two problems

ARUI - iOS AR Library

Published:

What is ARUI?

While building an Augmented Reality app in iOS, develoeprs need to spend time in handling all the ARkit specific details and sceneKit related code to setup an AR UI and then worry about the positioning and sizing the UI elements in AR. This library lets the developer specify a UI in the form of an Xib file and then parses the Xib file and creates an AR UI by resizing all the UI elements and arranging them to their relative positions in AR.

Arx Explore

Published:

What is Arx Explore?

Image Contrast Enhancement using Fuzzy Logic

Published:

The aim of the project is to perform efficient contrast enhancement with an improvement over traditional contranst enhancement techniques. We use around 7 fuzzy membership functions forming the Fuzzy Inference System which define the pixel manipulation. Since different fuzzy rules operate on different range of pixel values, contrast enhancement is better compared to techniques like Histogram Equalization. Histogram Equalization in our experiment on low resolution images, either over-enhances or under-enhances the images. The paper discusses the experiments in detail with results.