Publication

Abstract : Personality and attitude can be inferred from handwriting, a definite indicator of cognitive, emotional, and psychological traits. This paper suggests a handwriting-based psychological analysis method based on deep learning to classify depression levels and provide personalized recommendations to students. Three data sets are utilized in the study: 12-questions of Google Form survey on depression,anxiety,depression+anxiety,healthy corresponding handwriting samples, and stroke data of handwriting. EfficientNetB0 is utilized as a spatial pattern feature extractor in the suggested method. It has the highest accuracy of 81%. Depending on the classification, the system provides personalized strategies to improve mental health and study guidelines based on individual needs, based on professional judgments of medical doctors. The practice maximizes early detection of depression, promotion of mental well-being, and support to pupils in achieving maximum in academics.