TruEye: Fine-Grained Detection of AI-Generated Human Subjects in Images

2026-06-25

AI Summary

Existing methods for detecting AI-generated images suffer from poor generalization and limited interpretability, particularly struggling to perform fine-grained identification of synthesis types and effectively detect real-looking individuals embedded in authentic scenes. To address these limitations, this work proposes TruEye, a novel model that employs a mask-guided dual-stream Transformer to separately model human subjects and background scenes. By preserving spatial correspondence and leveraging region-gated cross-attention with token-level supervision, TruEye enables semantic consistency reasoning without relying on large language models. The method achieves the first fine-grained detection and localization across five distinct synthesis categories while offering interpretable predictions and significantly accelerating inference—up to 100× faster than LLM-based baselines. Extensive experiments on six public benchmarks and the newly introduced FineSyn dataset demonstrate that TruEye consistently outperforms state-of-the-art approaches in accuracy, generalization, and computational efficiency.

Abstract

AI generated images are proliferating across the Internet. While some are used for entertainment, others are weaponized for fraud and social engineering attacks on social media users. Existing detectors overfit to generators seen during training, treat detection as opaque binary classification, or rely on costly Large Language Models (LLMs) to explain their outputs. In this paper, we present TruEye, a novel model for fine grained detection and localization of AI manipulated or AI generated humans and scenes. Unlike conventional detectors that assign a single authenticity label, TruEye is the first to distinguish among five compositional categories of synthetic content, including the most challenging case in which a real human is composited into a real scene where they were never physically present. At its core is a mask conditioned dual stream transformer that separates human and scene tokens while preserving patch level spatial correspondence. Specialized reasoning within each stream and region gated cross attention enforce semantic coherence between subject and background, while token level supervision and global compositional classification yield robust, interpretable predictions without invoking an LLM. By restricting intra stream attention to semantically coherent tokens, TruEye also runs over $100\times$ faster than LLM based competitors. Experiments on 6 datasets and our newly curated FineSyn dataset, show that TruEye surpasses state of the art detectors with higher accuracy, faster inference, and stronger generalization to unseen AI generated or manipulated images.

Problem

• AI-generated images
• fine-grained detection
• image forgery
• human-scene composition
• synthetic content

Innovation

• fine-grained detection
• dual-stream transformer
• semantic coherence
• token-level supervision
• AI-generated image localization