AI-based Web App for Automated Direct Assessment of Learning Report Generation

In the academic evaluation process, manually entering marks from evaluated answer booklets into Excel sheets is time-consuming and error-prone. Accurate and efficient mark identification is crucial, as it directly affects student grade reporting. To address this issue, this project proposes a novel system for handwritten digit recognition leveraging OCR and CNN to generate Excel sheets with recognized digits. Existing OCR-based systems face challenges in handling diverse handwriting styles and require extensive pre-processing, leading to errors. The proposed system pre-processes images and utilizes CNN for feature extraction, predicting digits with high accuracy. The recognized digits are then organized into Excel sheets using the Openpyxl library these sheets provide structured data that can be directly used in Outcome-Based Education assessments and the generation of Faculty Course Assessment Reports. This facilitates more efficient analysis of student performance, identification of learning gaps, and implementation of targeted improvements, ultimately enhancing the reliability of the evaluation process