Answer sheet detection with image processing methods

Optical mark reading is a method of entering data into a computer system, marking the characters and markings made at defined positions on various forms or images. Thanks to this method, a large amount of data can be collected in a short time. It can be used on forms that people answer by markup or text, such as answers sheets, surveys and documents. Optical mark reading has become easier to use and can produce more comprehensive data with the development of image processing techniques today. In addition, handwritten data can be processed by recognizing characters through the convolutional neural network created by deep learning methods. In this developed project, the answer sheet layout is designed as frequently used, in accordance with the requirements and can be edited later. The regions to be marked on the answer sheet are grouped to eliminate data confusion. The system outputs the data as a result of the operations performed on the grouped fields using image processing methods and artificial neural networks on the answer sheet scanned to the system with the help of any scanner. This project basically realizes the information such as the number of questions requested from the user, leaving empty spaces and defining the fields to be filled with the help of image processing and deep learning instead of the classical optical reading process, minimizing the user dependency by detecting the information such as the software itself. In addition, by using the artificial neural network, it allows the entered characters to be read in classical exams by detecting them and develops a flexible reading without the need to introduce more editable forms. Keywords: Optical mark reading, Image processing, Deep learning, Convolutional neural networks