Can I read Deep Learning for Lip Reading Words on EtoBox?

Deep Learning for Lip Reading Words by Roopali Chavan is a document available to read on EtoBox.

What is Deep Learning for Lip Reading Words about?

This document presents research on classifying words from video data without audio using neural networks. It explores CNN+LSTM, Deep Layered CNN+LSTM, ImageNet-pretrained VGG-16+LSTM, and fine-tuned VGG-16+LSTM models on the MIRACL-V1 lip reading dataset. Data augmentation and techniques like dropout and batch normalization are used. The best model achieves 79% validation and 59% test accuracy for seen speakers, addressing a new area with applications in video analysis.

Author

Roopali Chavan

Language

EN