Integrating Bidirectional Long Short-Term Memory Networks with Convolutional Neural Network Models for Improved Translation Accuracy in Enhancing Gregg Shorthand Recognition
This study evaluates the performance of four CNN architectures (MobileNetV2, ResNet, DenseNet, and InceptionV3) integrated with BiLSTM for Gregg shorthand recognition. The best-performing model is implemented in G-Steno, an interactive application for shorthand practice and stroke assessment. The system delivers immediate feedback to learners, enhancing both the effectiveness and accessibility of shorthand learning.
Group Members
Karl Alexis R. RevillaJoshua R. AlbuferaDarren Jake A. AmbongJohn Paul L. BesagasMentor
Mr. Justine Jude PuraTopics
AgricultureAVP
Gallery
