About
I'am Vishnu
I am currently doing my Bachelors of Technology in Computer Science from Indian Institute of Technology, Mandi. I am interested in Machine Learning and Cybersecurity. I have a passion for coding and I am eager to innovate and create new solutions through my programming abilities.
Skills
Profesional Skills
Competent in C++, Java and Python Programming Languages. I have also worked with NodeJS, ExpressJS, Numpy, Pandas, Matplotlib, Streamlit.
Projects
AuthentiCheck
Engineered a high-accuracy real-time video deepfake detection system by integrating ViViT and Wav2vec models; with a detection precision of 85% for both pre recorded and live videos, minimizing security risks. Strengthened the solution with a user-friendly interface on Streamlit, enabling comprehensive video and audio feature fusion for robust deepfake identification.
GitHubStudent Management System
Provided complete management solutions to take care of all enrollment and maintenance reports (of a new or old student of the school). I had divided the program into small user-defined functions to take care of all future upgrades and expansion. Key Features of the project include user friendly menu driven options, exceptions used for database,tables, zero division to avoid inconvenience to users, Data Validation for functions (eg-sort(), delete(), etc.) to avoid ambiguous/wrong/invalid data entry.
Google DriveGuardians Of Truth
Implemented research paper using ProGAN dataset and CLIP (ViT-L/14) for feature extraction, achieving 92.9% and 87% accuracy on Dalle and Glide-100 datasets respectively surpassing SOTA results of 86% and 79%. Fine-tuned RESNET-50 model to distinguish real vs. fake images with 87% accuracy and detect DeepFake sources with 93% accuracy.Developed Guardian for Deepfake detection and TruthGuard for fake news detection, integrated into a web platform for real-time updates and verification, achieving 93% accuracy utilizing LSTM for sentiment analysis.
GitHub