Sharan Sahu's repositories
Direct-Image-Matching
Utilized OpenCV, ORBDescriptors, FLANN, Homography/Affine Transformations, and a multi-layer convolutional architecture to do direct image matching via feature and key-point matching for scale-variant images
Analyzing-The-Relationship-of-Spatial-Distribution-of-GHG-Emissions-and-Minority-Groups
ESPM 50AC Final Creative Project: Analyzing The Spatial Distribution of GHG Emissions and Minority Groups
calhacks-ml-model
Code for a differentially private Wasserstein GAN to create synthetic image and tabular data.
Citadel-Datathon-2021
Top 5 Project From Citadel Datathon 2021
CS-182-Project
WGAN Homework and Programming Assignment Creation For CS 182 Project
MLAutoFlow
MLAutoFlow is a package that allows users to push their custom-trained machine-learning models to Replicate without any installations or hassles. This tool is particularly useful for data scientists and developers who want to get their open-source machine learning models deployed fast without any hassles.
precision-recall-distributions
Assessing Generative Models via Precision and Recall (official repository)
PrivSynth
PrivSynth is a Streamlit application designed to create differentially private tabular data using Differentially Private Wasserstein Generative Adversarial Networks (DPWGAN). This tool is particularly useful for users who need to generate synthetic data sets that closely resemble original data while ensuring the privacy of individual data entries.
sharansahu
Config files for my GitHub profile.
sharansahu.github.io
A personal portfolio website created using HTML, CSS, JS, and AOS to store personal projects and papers to show to friends, family, and future employers.
stat453-deep-learning-ss21
STAT 453: Intro to Deep Learning @ UW-Madison (Spring 2021)
Surface-Deformity-Detection-With-CNNs
Machine learning methods and image processing were utilized to determine whether a surface contains a deformity. Through the development of a C++ program to generate surface deformities given image length, width, maximum deformity radius, and sample size as the training set, we utilize machine learning classifiers, namely Convolution Neural Networks (CNN), Support Vector Machines (SVC), and K-Means Clustering (KMC) to classify surfaces with either no impact, low impact, medium impact, or high impact.