PROJECT 01 / Applied AI · Media verification
TruthLens AI
A verification platform that brings image, video and news-article analysis into one workflow, with supporting evidence and uncertainty indicators.
01
The problem
Assessing potentially manipulated content requires different checks for different media. A single score can hide the evidence and uncertainty behind a result.
02
My approach
Built a centralized platform for analyzing images, videos and news articles, supporting three content types and more than seven media formats.
03
How it works
Combined metadata inspection, recompression analysis, video frame sampling, natural language processing and trained-model predictions to produce authenticity scores with supporting evidence.
04
What it demonstrates
Bringing multiple analysis methods into one application and presenting model output with evidence and uncertainty.
05
Limitations
Authenticity scores are assessments, not definitive proof. They need to be interpreted alongside the evidence and uncertainty shown. No benchmark accuracy or independent validation is claimed here.
KEEP EXPLORING
Loan Bias AuditorStock Market Analysis & PredictionSri Balaji Earth Movers