Abstract: At the present time evaluating an exam papers and declaring end result in a restricted period of time is a difficult task for educational schools, colleges, institutions, departments and Universities. Thus manual exam paper correction becomes more difficult and many fraudulent activities are happened today. To make it easier and more accurate the proposed aim is to develop a software for automatic exam paper evaluation and grading system, the system works by scanning the handwritten written exam papers then the scanned image be improved into an editable text using OCR tool and the evaluation will perform by matching the key terms which is maintained in the template. It is entirely integrated approach upon dissimilar level of knowledge by the method of examination, evaluation, result and formulation of subject papers. Code Shoppy In a field of education though teaching, evaluation and the performance method many organizations initiated with the use of advanced technologies. The approach of evaluating the examination papers is evolved although use of computer wherever the utilization of computer is obligatory in every varied techniques for analysis. This paper structured in different patterns such as offline, online and the manual exam paper evaluation with different analysis and techniques. Keywords— OCR; Ontology; Machine Learning; Evaluation System
To overcome the limitations of existing system, the proposed system aim is to build software for evaluating the handwritten mathematical examination papers and the theoretical papers automatically. In this system, the handwritten examination papers are scanned and converted into an editable format using OCR tool and then evaluation will be performed by formulas, in-between steps and final solution for the numerical papers, for the theory papers the evaluation will be performed by the keywords or synonym based keywords which are maintained in the database. Finally the marks will be given depend upon the identical key terms. Fig.3.1 System ArchitectureA. OCR Tool: Optical Character Recognition tool that permits to convert the dissimilar types of scanned paper documents, captured images by digital camera upon an editable form or PDF files In order to repurpose and extract data from camera images, scanned documents or PDFs, the OCR software would particular exposed letters on image and locates them into words then words to sentences by enable to edit and access the original document content. B. Binarization Binarization converts a picture from grayscale or color keen on black along with white known as a “binary image” as present is two colors. It also converts the acquire form of image into binary format, inside which foreground contain symbol, filled data, frame line and printed entities. C. Scanning Towards the model selection for system process a figure recognition step match the features which extracted from an incoming figure next to where the extraction from each design of a module D. Image extraction The extraction from relevant data though particular field and preprocesses the information in order to enhance data and eliminate noise. The characters are recognized by the extraction of field is nearby subsequent to the present analysis. E. Pre-processing Stage Pre-processing step involves RGB to grayscale image alteration, noise removal, Thresholding etc. In RGB to grayscale picture conversion, RGB image is converted to gray image. Thresholding converts gray to black and white image. If needed the key document could be resized if needed. This image pre-processing step has immense impact on quality of OCR process. F. Smoothing Smoothing operations are used for noise decrease and for blurring in color images. Blurring used inside the preprocessing stage for as removal of undersized details from image. Smoothing operations reduce the noise in binary image. G. Cleaning Cleaning is done for the separation and removal of preprinted entity whereas the preserving of filled data by resources of image-processing technique. The appearance of summarize information commencing the knowledge base for the utilization of cleaning procedure.H. Enhancing It reconstructs the stroke that has been detached throughout cleaning process. Over the baseline if some filled data is printed then in cleaning it also remove the foremost discontinuity in character which might obstruct the recognition.I. Character Recognition Recognition is to map the given pattern with internally stored database. The matching though the recognition process is presented by the gesture with the standard indication. The important features like the size, depth, shape, color are extracted from the recognition system by input image. If the geometrical features are extract as of input after that the character is able to be coordinated by means of normal characters within the library. Thinning process will remove the pixels consequently the object with no holes shrink toward a minimally associated stroke. Skeletons are useful for describing properties of a shape and also in various cases it used intended for transformation of the original shape.
The modules of the proposed work mainly focus on conversion part, whereas handwritten recognition remains a challenging one while compare toward further assessment process. The proposed system for automatic exam paper correction is build till now the extraction. It contains the following: 1) Student Login: The student login created for the students to view their individual results. The students are able to view results by the valid register number and the password also they can view the marks scored for particular answer. 2) Admin login: Admin login will be controlled by the administration. This will set by entering the abuser given name and code word. Here the Admin be able to edit or create examination paper patterns. The evaluation will perform by the administration though the keys which maintained in the database. They will perform the evaluation by extract summary, extract calculation and report detail for both the theoretical and numerical papers. Fig.4.1 Conversion of Handwritten notes to Text Fig.4.1 shows the Conversion of Handwritten notes to Text from the uploaded image Fig.4.2 Admin Home page Fig.4.2 shows the admin home page of extraction part by entering the user name and password. Fig.4.3 Summary Extraction Fig.4.3 shows the summary extraction from the uploaded image. Fig.4.4 Calculation Extraction Fig.4.4 shows the calculation extraction from the uploaded imageFig.4.5 Admin Report Fig.4.5 shows the Admin Report and the reports can be updated or deleted.
The proposed work is completed till this, the system also provide grade depend upon the mark which scored by the individual student. Here the marks will be awarded by matching key terms which maintained in the database and the results of the students are automatically stored into the database. Correction Mode Accuracy Error Rate Time Manual paper correction > 80% < 20% > 4 min Automatic paper correction > 98% < 2% < 1 min Table.1.1 Performance Metrics
Technology enrichment within constituent component hardware and software technology, assessment pattern has spectacularly changed as of class base exam to an online assessment. Online automated examinations do not just implement while preset examination and assessment process although provide reliable and quick appraisal with more flexibility for computerization of whole assessment process. These techniques would encompass their individual advantages consequently the dissimilar type of automatic evaluation is important. Finally the paper is finished with most recent improved functioning of automatic evaluation process. Use of computer have significantly increased in education learning practice and approximately each pasture of education as well extensively use for conduct, evaluate and display consequence of such assessment. Development and effectiveness can exist obtain more time towards the most recent needs, thus the system is easier and more accurate.