Handwritten Prescription Deciphering using CNN and Bi-LSTM

• Developed a hybrid deep learning model combining CNN for spatial feature extraction and Bi-LSTM for sequential dependency reduction.
• Achieved a high recognition accuracy of 98.40% on test data.
• Implemented an iterative training methodology to prevent overfitting and ensure continuous learning.
• Contributed to improved patient care and healthcare information management by reducing misinterpretation of handwritten prescriptions.

QuizNotepad

• Created an engaging quiz game featuring over 10 coding language puzzles, showcasing interactive and immersive design.
• Seamlessly integrated the game with a backend system, resulting in a 30% boost in user engagement and interaction.
• Spearheaded the development of a feature-rich environment, enabling users to create, save, and revisit over 20 personal notes, showcasing innovative and user-centric design.

Pneumonia Detection using Machine Learning

• Developed a high-performance sequential model for Pneumonia detection, achieving an impressive accuracy rate of 94.5%, demonstrating exceptional proficiency in machine learning.
• The management of a sizable 2 GB dataset showcased adept data handling and analysis skills, contributing to the achievement of a robust 96% confidence level for accurate diagnostics.