Text Extraction and Audio Generation from Images using NLP
Ravuru Rakesh, Dwarakanath, Malathi Eswaran, Arunkumar Chinnasamy, Chetaly Mawal · 2023
By producing a digital replica of a physical object on devices, digitization enables us to immortalize that object. It saves us time by eliminating the need to manually search through tangible storage devices like albums and notebooks, and it gives us access to tools for managing and securing data. To keep a record of the information on receipts or invoices, identity cards, and nutritional labels, pictures of it are taken frequently. Automating the information extraction and documentation processes will advance this. Now, one has the knowledge necessary to develop text extraction and detection technologies with the help of developments in computer vision. But it is still a problem because documents with chaotic layouts, blurry images, and noise around the text produce very poor text accuracy and extraction outcomes. In order to overcome this obstacle, the image would need to be significantly improved using pre-processing methods as Brightness Correction, Contour Detection, Skewness Correction, Morphology, and Binarization. Prior to text extraction, a mechanism created using the greatest possible mix of image pre-processing techniques can significantly increase text accuracy.