Analysis and recognition of facial expression is an important aspect in the field of research and so it is required to achieve optimum recognition rate. Currently research issues are to pay emphasis to increase the recognition rate which can be done by refining the pre-processing of datasets, the method for extracting the features and using the best classifier for emotion recognition. Recognition rate for facial emotion recognition depends on the important step of feature extraction. To improve the recognition rate, if features are extracted using different ways or forecast then most likely the laying-off will rise which may result in fall in the recognition rate. The main issue with Gabor filter is high dimension and high redundancy although it has extreme alteration of features. Dimension and redundancy should be reduced using some method. The dimension reduction method for Gabor is called filtering so this whole system is called Gabor filter. These filtering method are sampling, averaging and PCA etc. In the projected Gabor feature removal method, wavelet transformation is used for filtering the Gabor features and it obtains finest features from facial Gabor matrices.

Keywords: DWT, Expression, Facial Emotion Recognition, Gesture, Gabor Filter.