GokulVSD
About Blog Projects Awards Resume
Crypticket
| Crypticket

An offline capable cryptographic ticket generation and authentication platform using service workers and local storage caching. Built as a responsive PWA from scratch using React, utilizing EdDSA elliptic curve cryptography

ReactPWACryptographyCSSJavaScriptjQueryBootstrapResponsive DesignLocal Storage APIBrowser CachingReactToolTipsEdDSAOffline CapableElliptic

MonoDAC
| MonoDAC

Monocular depth estimation built with OpenCV and TensorFlow using an encoder-decoder, employing Atrous Convolutions and ASPP, along with a web UI for real-time wireless image capture, depth inference and 3D point clouds

Deep LearningComputer VisionDepth EstimationDeepLabv3+3D Point CloudIPcamera

Wuasta
| Wuasta

A predictive alarm Android app built in Kotlin that pragmatically rings at the right time, taking into account real-time traffic conditions and historical data to find when to depart in order to arrive at another location at a predefined time

Smart AlarmTrafficAndroidUIDistance Matrix APIPlace Picker API

| CineManagerDB

A comprehensive theatre management and ticket booking system built with Flask for Python and MySQL, presented via a web based UI.

FlaskMySQLPythonJavaScriptCSSDatabaseTicket Booking

| tiCtactoe

Tic Tac Toe game with GUI. PvP as well as PvC with rudimentary AI. Written in C using GTK lib 3.0 and Glade.

GTKCTic Tac ToeGlade

| UnicodeBankDB

A file based bank database management web app written in Java using Spark and Maven, with indexed variable length nested records, concurrency control, state transfer and Huffman Compression.

Java SparkHuffman CompressionText DatabaseJavaDigital BankCSSJavaScript

| YTrendNet

Dual Implementation of Artificial Neural Networks to find how long a YouTube Video will stay trending.

KerasTensorFlowArtificial Neural NetworksYouTubeNumPyPython

| MidLLaMAI

Benchmarking suite for evaluating performance of compressed Hugging Face LLaMA models, focused on finding an optimal balance between different pruning and quantization techniques with varying number of LLM model parameters

QuantizationPruningLLMBenchmarkingGPTQAWQ

| SymptomJournal

An Android app that measures vitals using the breathing rate using an accelerometer placed on your torso, and your heart rate using the camera and flash to measure the change in color intensity of your blood

Mobile ComputingAndroidKotlin

| TicTacTmObilE

Tic-Tac-Toe game on Android that allows an AI to play against a human opponent. The primary goal is to create a game where the computer plays against a human player using the Minimax algorithm with alpha-beta pruning

Mobile ComputingAndroidKotlinBluetoothJetpack ComposeRoom

| ViVeC

Efficient search and retrieval of relevant images from the Caltech-101 dataset using distance/similarty functions between feature descriptors generated using Color moments, Histograms of oriented Gradients (HoG) and ResNet layers

PythonCaltec101Color MomentsHoGResNet

| ViVeC2

Research and implement dimensionality reduction techniques such as SVD, LDA, K-Means, CP Decomposition & Non-Negative Matrix Factorization to generate latent semantics in order to provide the top-k similar images or labels based on given images or labels for Caltec101. We include personalized Page Rank algorithm for similarity propagation

PythonCaltec101SVDLDAK-MeansCP DecompositionNNMFPageRank

| ViVeC3

Using latent semantics, we build clustering, classification, and approximate multidimensional indices for efficient information retrieval on the Caltec101 Dataset. Includes efficacy with precision/recall, and Relevance Feedback using SVM and Probabilistic Models

PythonCaltec101DBScanLSHSVMRelevance FeedbackClassificationHashing

by Gokul Vasudeva | GokulVSD