Technology Category
Computer Vision: Generative Adversarial Networks (GANs) for VisionMachine Learning: Multimodal LearningIntelligent Robots: Multimodal Perception & Sensor Fusion
Handcrafted Feature Fusion for Reliable Detection of AI-Generated Images
🤖 AI Summary
This study addresses the pressing challenge posed by generative AI–synthesized images to the authenticity of digital media by proposing a detection method based on multimodal handcrafted features and ensemble learning. The discriminative efficacy of features—including DCT, HOG, LBP, GLCM, wavelet transforms, and color histograms—is systematically evaluated on the CIFAKE dataset, and multiple feature sets are fused as input to ensemble models such as LightGBM, XGBoost, and CatBoost. Experimental results demonstrate that feature fusion substantially enhances detection performance. Notably, LightGBM with the combined feature set achieves a PR-AUC of 0.9879, ROC-AUC of 0.9878, F1 score of 0.9447, and Brier score of 0.0414, outperforming single-feature approaches while maintaining high accuracy, interpretability, and computational efficiency.
Technology Category
Application Category
📝 Abstract
The rapid progress of generative models has enabled the creation of highly realistic synthetic images, raising concerns about authenticity and trust in digital media. Detecting such fake content reliably is an urgent challenge. While deep learning approaches dominate current literature, handcrafted features remain attractive for their interpretability, efficiency, and generalizability. In this paper, we conduct a systematic evaluation of handcrafted descriptors, including raw pixels, color histograms, Discrete Cosine Transform (DCT), Histogram of Oriented Gradients (HOG), Local Binary Patterns (LBP), Gray-Level Co-occurrence Matrix (GLCM), and wavelet features, on the CIFAKE dataset of real versus synthetic images. Using 50,000 training and 10,000 test samples, we benchmark seven classifiers ranging from Logistic Regression to advanced gradient-boosted ensembles (LightGBM, XGBoost, CatBoost). Results demonstrate that LightGBM consistently outperforms alternatives, achieving PR-AUC 0.9879, ROC-AUC 0.9878, F1 0.9447, and a Brier score of 0.0414 with mixed features, representing strong gains in calibration and discrimination over simpler descriptors. Across three configurations (baseline, advanced, mixed), performance improves monotonically, confirming that combining diverse handcrafted features yields substantial benefit. These findings highlight the continued relevance of carefully engineered features and ensemble learning for detecting synthetic images, particularly in contexts where interpretability and computational efficiency are critical.
Problem
Research questions and friction points this paper is trying to address.
AI-generated images
image forgery detection
handcrafted features
synthetic image detection
media authenticity
Innovation
Methods, ideas, or system contributions that make the work stand out.
Handcrafted Features
Feature Fusion
AI-Generated Image Detection
LightGBM
Interpretability
🔎 Similar Papers
No similar papers found.
