Many multimedia applications and entertainment industry products like games, cartoons and film dubbing require speech driven face animation and audio-video synchronization. Only Automatic Speech Recognition system (ASR) does not give good results in noisy environment. Audio Visual Speech Recognition system plays vital role in such harsh environment as it uses both -audio and visual -information. In this paper, we have proposed a novel approach with enhanced performance over traditional methods that have been reported so far. Our algorithm works on the bases of acoustic and visual parameters to achieve better results. We have tested our system for English language using LPC, MFCC and PLP parameters of the speech. Lip parameters like lip width, lip height etc are extracted from the video and these both acoustic and visual parameters are used to train systems like Artificial Neural Network (ANN), Vector Quantization (VQ), Dynamic Time Warping (DTW), Support Vector Machine (SVM). We have employed neural network in our research work with LPC, MFCC and PLP parameters. Results show that our system is giving very good response against tested vowels.
In this paper, we propose a novel appearance based local feature descriptor called Local Mean Binary Pattern (LMBP) for facial expression recognition. It efficiently encodes the local texture and global shape of the face. LMBP code of a pixel is produced by weighting the thresholded neighbor intensity values with respect to mean of 3 3 patch. LMBP produces highly discriminative code compared to other state of the art methods. The micro pattern is derived by thesholding on mean of the patch, and hence it is robust against illumination and noise variations. An image is divided into M N regions and feature descriptor is derived by concatenating LMBP distribution of each region. We also propose a novel template matching strategy called Histogram Normalized Absolute Difference (HNAD) for comparing LMBP histograms. Rigorous experiments prove the effectiveness and robustness of LMBP operator. Experiments also prove the superiority of HNAD measure over well-known template matching methods such as L2 norm and Chi-Square. We also investigated LMBP for expression recognition in low resolution. The performance of the proposed approach is tested on well-known datasets CK, JAFFE, and TFEID.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.