FAAVSP - The 1st Joint Conference on
Facial Analysis, Animation, and
A study is presented to determine the relative importance of different visual features for speech recognition which includes pixel-based, model-based, contour-based and physical features. Analysis to determine the discriminability of features is performed through F-ratio and J-measures for both static and temporal derivatives, the results of which were found to correlate highly with speech recognition accuracy (r=0.97). Principal component analysis is then used to combine all visual features into a single feature vector, of which further analysis is performed on the resulting basis functions. An optimal feature vector is obtained which outperforms the best individual feature (AAM) with 93.5% word accuracy. Index Terms: Visual features, speech recognition, F-ratio, J-measure, PCA
Bibliographic reference. Websdale, Danny / Milner, Ben (2015): "Analysing the importance of different visual feature coefficients", In FAAVSP-2015, 137-142.