Analysis of Image Classification for Text Extraction from Bills and Invoices

Optical Character Recognition (OCR) technology offers a complete alphanumeric recognition of printed or handwritten characters from pictures such as scanned bills and invoices. Intelligent extraction and storage of text in structured document serves document analytics. The current research attempts to find a methodology through which any information from the printed invoice can be extricated. The intermediate image is passed over using an OCR engine for further processing. Segmentation extracts written text in various fonts and languages. Image classification helps in making a decision based on the classification results. This paper surveys these techniques and compares them in terms of metrics, algorithm and results.