TEXT AND IMAGE PLAGIARISM DETECTION USING NLTK

Abstract

The abundance of content across academic, media, and online platforms in the digital age has raised the possibility of plagiarism in both written and graphical forms. Using the Natural Language Toolkit (NLTK) for text analysis and traditional image detection, this study suggests a hybrid method for detecting plagiarism in both text and images. processing techniques for identifying visual similarity. Tokenization, elimination of stop words, stemming, and semantic similarity calculation using cosine similarity and the Jaccard index were all performed for textual data using the NLTK package. Simultaneously, feature extraction methods including histogram comparison and perceptual hashing (pHash) were used to detect image plagiarism by identifying modified or repeated images. The integrated technology highlights reused or slightly modified pictures in addition to identifying plagiarism in both exact and paraphrased text. The results show that the combined model provides high detection accuracy. different types of plagiarism, which makes it a useful resource for publishers, educational organizations, and content verification services. This research emphasizes how important it is to combining computer vision and linguistic processing to create a more thorough plagiarism detection system.