/AIorReality

Image Classifier for Recognizing AI-Generated Images Based on SVM

Primary LanguagePython

AIorReality: Image Classifier for Recognizing AI-Generated Images Based on SVM

demo of real-time detection

demo.mp4

Source Code location Guide:

To facilitate users, we provide a brief guide that users can use to identify the location of the given part of the source code in AIorReality. The table is as follows:

Table 1: Source Code Location Guide for AIorReality
Main Directory Subdirectory Content
datasets AI_from_kaggle Store AiArtData and RealArt
model_saved -- Pre-trained models are used for real-time detection
Result Grid_search Stores the results of the grid search, including logs and heat maps
Train Stores the results of model training, including logs, loss trends, and confusion matrices
temp_picture -- Used to store the images captured in the video and predict the results of the pictures
test -- Used to get footage for still video
util image_to_video.py Generate a still video from the image
logo.ico UI logo
output.mp4 Generated static video
rename.py Used for recoding data sets
Classifier.py -- It is used for classifier construction and model training
Feature_extraction.py -- 4 ways of feature extraction
Grid_search.py -- Perform a grid search to obtain optimal hyperparameters
main.py -- UI interface: Used for visual classification effect display