In this study, a fast, reliable and cheap method is suggested for the recognition and evaluation of the marks of a multiple-choice test on the images that are obtained via the scanning of the optical forms printed on a standard sheet of paper with an ordinary scanner. This method is called the recognition of optical marks and is the process of capturing the data on the multiple-choice forms. The application of recognition has been developed by using the software language, Python and the image processing library, OpenCV. When the answer sheet is loaded in the application, incorrect answers are marked as red while the correct ones as green and with the calculation of the correct/incorrect answers and blanks, the result is printed on the optical form image. This method is economical, fast and quite successful. As a result of three examinations in Konya with the participation of 35.250 students, 105.750 optical forms were scanned with a scanner. The recognition success has been calculated as 99,76 %. The empirical studies have shown that the suggested system is more successful than the conventional optical mark recognition systems regarding accuracy, reliability and performance.