HN Showcase
Back to browse
FakeSec – AI photo detector with 12 independent analysis methods

FakeSec – AI photo detector with 12 independent analysis methods

by theevoq·Feb 20, 2026·1 point·1 comment
Visit ProjectView on HN

AI Analysis

●●SolidBig BrainEye Candy

12-method fusion (ELA, FFT, noise, pixel analysis) beats single-model detectors, but detection arms race is endless.

Strengths
  • •Genuine technical depth: ELA, FFT, DCT, Benford's Law, and pixel-level forensics—not just one neural net.
  • •Explainable scoring per method; users see which signals triggered the verdict.
  • •Polished landing page with live demo, clear pricing tiers, Telegram bot integration.
Weaknesses
  • •Detection quality ultimately depends on training data; adversarial images will defeat any fixed method set.
  • •Crowded category—remove.bg clones, Sensetime DeepFake, other forensic tools already exist; no moat if adversaries keep iterating.
Category
AI/ML
Target Audience

Content moderation teams, journalists, media verification professionals, casual users checking photos

Similar To

Sensetime DeepFake Detection · Reality Defender · Truepic forensic verification

Similar Projects

Health●●Solid

I built an AI trainer and calorie scanner

Photo-to-macros logging and a context-aware workout engine are the project's real selling points — snapping a meal instead of digging through a database and having workouts adapt to your current equipment are tangible time-savers. The landing page is clean and trustworthy, but there’s no evidence on the page about CV accuracy, model safety, or sample workout flows; those are the two things that will make or break this in practice.

Solve My ProblemShip It
mihaibundea
103mo ago