Seeing Can
Still Be
Believing.
eWitness uses SmartHash — a peer-reviewed DCT-based perceptual hashing algorithm — to establish tamper-proof provenance for every photo and video you capture.
Built on Trust,
Powered by Truth
SmartHash-Powered
eWitness uses SmartHash — a DCT-based perceptual hashing algorithm developed at John Jay College of Criminal Justice — to create a unique fingerprint of every photo and video at the moment of capture.
Privacy First
eWitness never stores your personal information. Your media stays on your device until you choose to share it. Only a blockchain transaction ID is embedded in the file metadata.
Tamper-Proof Authentication
Any content-changing modification to your media after capture is immediately detectable. SmartHash distinguishes between innocent edits (compression, brightness) and malicious tampering (object addition or removal).
Designed for Smartphones
SmartHash is fast and efficient — engineered to run in real time on resource-constrained mobile devices. Download from Google Play and authenticate media with zero friction.
Publicly Verifiable
Anyone can verify the authenticity of your media by looking up the blockchain transaction ID embedded in the file. No special software or accounts required.
Forensic-Grade Trust
Built on peer-reviewed research published at ACM CIKM 2024, eWitness meets the evidentiary standards demanded by legal proceedings, investigative journalism, and institutional research.
Truth Matters
Everywhere
Citizen Journalism
Record events with cryptographic proof that footage is unaltered. Protect the integrity of your reporting.
Photography Rights
Establish undeniable ownership of your creative work. Track usage and ensure proper attribution across the web.
Legal Evidence
Capture incidents with a timestamped, tamper-proof record suitable for legal proceedings and insurance claims.
Research & Academia
Document fieldwork with verifiable provenance for peer review, institutional records, and fact-checking workflows.
SmartHash: Perceptual Hashing for Image Tampering Detection and Authentication
ACM CIKM '24 — 33rd International Conference on Information and Knowledge Management · Boise, ID · October 2024
SmartHash is evaluated on 5 large tampered-image datasets (CASIA2, PhotoshopBattle, IMD2020, DEFACTO, MSCOCO) and outperforms Apple's NeuralHash and Microsoft's PhotoDNA for image authentication tasks, achieving up to 99.45% accuracy — while running 6× faster than NeuralHash on a single CPU core.
Read the Paper →The Smart Contract
Behind Every Proof
Every SmartHash registered by eWitness is recorded as a permanent, immutable Set Metadata transaction on a public blockchain contract. Anyone can inspect it — no login, no trust required.
Start Proving Truth Today
Join photographers, journalists, researchers, and everyday citizens who trust eWitness to authenticate their media.