Microscope patent illustration
INTEGRITY

cnn based ai forensic system

analyze image
Technical component diagram

Overview

Integrity is a forensic AI detection tool that uses a multi-branch Convolutional Neural Network to determine whether an image is real or AI-generated. It analyzes images across three layers: spatial patterns, frequency anomalies, and embedded metadata (EXIF and C2PA tags), fusing these signals to deliver a confident verdict. Integrity can detect AI-generated faces, unusual lighting artifacts, and fabricated metadata injected by image generators. Designed for speed and accuracy, it offers both a quick scan using our lightweight custom CNN and a deep scan powered by our deep custom CNN, making digital media verification accessible to everyone.

upload

click to launch analyzer

(interactive demo)

select model

how to use

QUICK SCANlightweight

Fast, lightweight model designed for rapid preliminary scanning with minimal computing overhead.

DEEP SCANthorough & heavier

Deep custom CNN forensic model delivering superior accuracy and multi-layer feature inspection.

ILLUSTRATION DISCLAIMER

Integrity currently does not support digital illustrations, artwork, or sketches. If the gatekeeper model determines an upload is an illustration and cannot provide a reliable verdict, the analysis will halt immediately.