Extraction of Logo from Real Time Document Images Using Masking and Median Filter Approaches

In this paper, we propose an efficient method for extracting the logo from real-time document images. A logo represent key perspectives and thoughts in a specific item, making it the most important choice for brands with a lot to say. These logos can include line-drawn symbols, shapes, patterns, or illustrations that use non-literal illustrations to communicate brand awareness. Frequently combined with a logotype. The proposed method extract the logo based on median filter and masking techniques. Experimentation is conducted with two stages namely logo extraction with enhancement and logo extraction without enhancement. Experimentation is carried out on our own data set containing 500 document images and have obtained 90% accuracy for logo extraction with enhancement and 86% for without enhancement.