Int. J. Emerg. Sci., 5(2), 68-75, June 2015 ISSN: 2222-4254 © IJES

Facial Expression Analysis in Video Using Discrete Wavelet Transform D. Magdalene Delighta Angeline1, I. Samuel Peter James2 1

Department of Computer Science and Engineering, Dr.G.U.Pope College of Engineering, Sawyerpuram-62825, Tamilnadu, India., 2 Department of Computer Science and Engineering, Chandy College of Engineering,Thoothukudi, Tamilnadu, India., 1

[email protected], [email protected]

Abstract. Face recognition technology has evolved as a popular identification technique to perform verification of human identity. By using the feature extraction methods and dimensionality reduction techniques in the pattern recognition applications, a number of facial recognition systems have been produced with distinct measure of success. This paper proposes a novel technique for eye detection using color and morphological image processing. It is observed that eye regions in an image are characterized by low illumination, high density edges and high contrast as compared to other parts of the face. The method proposed is based on assumption that a frontal face image (full frontal). This paper provides a research on an efficient illumination invariant face recognition system using Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA). This technique is found to be highly efficient and accurate for detecting eyes in frontal face images. Keywords. Discrete Wavelet Transform, Feature Extraction, Principal Component Analysis, Skin Tone

1.

INTRODUCTION

The human face provides one of the most powerful, versatile and natural means of communication, during social interaction. Human beings express emotions in day to day interactions Understanding emotions and knowing how to react to people’s expressions greatly enriches the interaction According to Meharabian [1], Facial Expressions provide important communicative cues, which constitute 55 percent of the effect of a communicated message; hence recognition of facial expressions became a major modality in Human Computer Interaction. The possibility of making Computers to recognize facial expressions and use the information in Human Computer Interaction has gained significant research interest over the last few years. This has given rise to a number of automatic methods to recognize facial expressions in images or video [2-7]. Mehrabian [1] indicated that only 7% of message is due to linguistic language, 38% is due to paralanguage and 55 % of message is communicated by facial expressions. This implies that the facial expression is a major modality in human face-to-face communication. Thus we can imagine that, when designing the Human Computer Interfaces (HCI), the facial expressions seems to be a major factor for improving the communicability of message, even in human-machine communication.

68

D. Magdalene Delighta Angeline and I. Samuel Peter James

2.

LITERATURE REVIEW

Facial expressions are not only emotions, but other mental activities, social interaction and physiological signals. In this survey, we introduce the most prominent automatic facial expression analysis methods and systems presented in the literature. Facial motion and deformation extraction approaches as well as classification methods are discussed with respect to issues such as face normalization, facial expression dynamics and facial expression intensity, but also with regard to their robustness towards environmental changes [8]. The approaches to facial expression recognition can be, roughly divided into two classes: geometrical feature-based approaches and appearance-based approaches [9]. The geometrical feature-based approaches rely on the geometric facial features, which represent the shapes and locations of facial components such as eyebrows, eyes, nose, mouth etc. As for the appearancebased approaches, the whole-face or specific regions in a face image are used for the feature extraction via optical flow or some kinds of filters. Many of the previous works [10, 11] on facial expression recognition are based on the existence of six universal expressions (anger, disgust, fear, joy, sorrow and surprise). Conventional methods extract features of facial organs, such as eyes and a mouth, in gray or color images of front faces and recognize the expressions from changes in their shapes or their geometrical relationships by different facial expressions [6, 12-14]. However, estimation of their precise positions and shape attributes in real images is difficult, because of the wide variety of the face features, skin color/brightness, illumination conditions and geometrical variations such as head orientations. As a result, many of the systems need human assistance such as attaching marks on the subject’s face or specifying windows covering each organ in the image. The success of facial expressions recognition system depends heavily on how well the movement of the key features points, like eyeballs and mouth corner, are tracked on the human face. Facial expressions do not actually involve the entire face, but rather specific sets of muscles under the face near the eyes, nose and mouth.

3.

METHODOLOGY

In the proposed method, facial expressions of the human face are identified from the input image using Eigen spaces method. If the input image is similar to some expression training set, the reconstructed image will have less distortion than the image reconstructed from other eigenvectors of training expressions. With this idea, the training set is divided into six classes according to universal expressions and computed the Eigen spaces of each class. For a test face image, we first project it onto the Eigen space of each class independently and then derive reconstructed image from each Eigen space. By measuring the similarity between input image and the reconstructed image of each class, we can identify the class of input image whose reconstructed image is most similar to the input one. 3.1 Discrete Wavelet Transform algorithm

In these applications one needs a fast and memory efficient algorithm to compute the transform. In traditional n-D DWT algorithms, Multi-dimensional discrete wavelet transform has been considered to be used in many fields such as image and video processing applications. In order to employ the advantages of our algorithm for multi-dimensional signals such as video, it is proper to use its structure only for the longest dimension. Then the procedure outputs the prepared 69

International Journal of Emerging Sciences 5(2), 68-75, June 2015

coefficient and inputs the next sample to the buffer. The wavelet transform has been successfully applied to image and signal processing. Unlike the Fourier transforms, the wavelet transform can detect the characteristics of local signals. The wavelet transform is particularly helpful for the analysis and processing of interhistograms where local fringe information is important. Research on the wavelet transform for interferogram processing has been active for a decade, and different schemes have been proposed. Wavelet transform decomposes a signal into a set of basis functions. These basis functions are called wavelets. Wavelets are obtained from a single prototype wavelet y(t) called mother wavelet by dilations and shifting:

 a ,b (t ) 

1 t b ( ) a a

(1)

Where a is the scaling parameter and b is the shifting parameter. The wavelet transform is computed separately for different segments of the time-domain signal at different frequencies. Multi-resolution analysis: analyzes the signal at different frequencies giving different resolutions. MRA is designed to give good time resolution and poor frequency resolution at high frequencies and good frequency resolution and poor time resolution at low frequencies. Good for signal having high frequency components for short durations and low frequency components for long duration e.g. Images and video frames. The 1-D wavelet transform is given by: (2)

The inverse 1-D wavelet transform is given by: (3)

Discrete wavelet transform (DWT), which transforms a discrete time signal to a discrete wavelet representation. It converts an input series x0, x1, ..xm, into one high-pass wavelet coefficient series and one low-pass wavelet coefficient series (of length n/2 each) given by:

(4)

(5)

where sm(Z) and tm(Z) are called wavelet filters, K is the length of the filter, and i=0, ..., [n/2]1. In practice, such transformation will be applied recursively on the low-pass series until the desired number of iterations is reached. 3.2 Principal Component Analysis

Principal Component Analysis (PCA) is a dimensionality reduction technique which is used for compression and recognition problems. It is also known as Eigen space Projection or KarhunenLoeve Transformation. PCA projects images into a subspace such that the first orthogonal 70

D. Magdalene Delighta Angeline and I. Samuel Peter James

dimension of this subspace captures the greatest amount of variance among the images and the last dimension of this subspace captures the least amount of variance among the images. The main goal of PCA is the dimensionality reduction, therefore the eigenvectors of the covariance matrix should be found in order to reach the solution. The eigenvectors correspond to the directions of the principal components of the original data, their statistical significance is given by their corresponding eigen values. PCA tries to identify the directions of maximum variation in the training data. To identify an input image we proceed as follows:  Evaluate the components of input image along the selected k eigen vectors.  Reconstruct the image from the components.  If the distance between the reconstructed image and the original image is above a threshold, the input image is not a face image.  Compute the distance of the input image from the training images in the space spanned by the k eigen vectors. If the minimum distance is above threshold  then the input image is not a face from the training database else report the training image with the minimum distance as the recognized image. The threshold  is defined as  = ½ max {||dij||} (6) where dij denotes the distance between training images i and j. The steps in finding the principal components are summarized as follows:  Minimize representational error in lower dimensional subspace of input  Choose eigenvectors corresponding to m largest eigen values of total scatter as the weight matrix (7) (8)

Input N Faces

Normalization n x m

PCA (N Eigen Faces, Each n⋅m values)

Choose M first Eigenfaces According to corresponding Eigen values

Form vectors from pixels

Create matrix for N vectors

Project each face into N eigen face

N Vectors

71

International Journal of Emerging Sciences 5(2), 68-75, June 2015

Figure 1: PCA Feature Extraction

3.3 Algorithm for Edge Detection

Edge detection is one of the Emotion algorithms used in video segmentation and Human Emotion. Many edge detection algorithms have been proposed. Demigny proposed a new organization of filter at 2D and 1D levels which reduce the memory size and the computation cost by a factor of two for both software and hardware Implementations. An Edge detection algorithm for eye detection on face images, including low resolution ones, is presented in this paper. Pixel intensity information might prove unreliable due to varying lighting conditions. It is main purpose low resolution images; details of the eyes sometimes are lost. Thus, an approach detects the eyes region on a face. Detection Algorithm consists of a feature finder and a face morphed. 3.4 Skin Tone Algorithm

Detecting human skin tone is one of most important numerous applications such as video surveillance face and image recognition, human and computer interactions. And it’s mainly used to image retrieval and image editing and stenography concept using hiding information’s. The histogram of skin pixels and non-skin pixels of the present frame of and skin area and non-skin area computed in the image. Modeling skin color implies the identification of a suitable color space and the careful setting of rules for cropping clusters associated with skin color. 3.5 Face Recognition

Detecting and analyzing a face in a single image is the most important part of almost all face recognition systems. Face localization aims to determine the image position of a face for verification purpose of documents such as passport, driving license, ID cards, etc. Face recognition has attracted much attention. Many research demonstrations and commercial applications have been developed from these efforts. A first step of any face processing system is detecting the location of faces in images. This task however, is complicated because of orientation, scaling, rotation, illumination, etc. which are present in most images. These rulebased methods encode human knowledge of what constitutes a typical face. Usually, the rules capture the relationships between facial features. These methods are designed mainly for face localization. These algorithms aim to find structural features that exist even when the pose, viewpoint, or lighting conditions vary, and then use these to locate faces. These methods are designed mainly for face localization. In face detection and recognition area we are interested in high information areas such as eyes, nose, mouth etc. These areas convey fundamental information needed for face recognition.

4. RESULTS AND DISCUSSION The result utilizes utterances that have at least one label from the target emotional set and to identify at least one of the emotions reported by the evaluators. This problem is difficult because the classifiers must be able to identify the categorical emotion labels when the evaluators 72

D. Magdalene Delighta Angeline and I. Samuel Peter James

themselves could not. The results demonstrate that the emotional profiling technique is able to effectively capture the emotion content inherent even in ambiguous utterances. 4.1 Performance

Using DWT, Peak Signal to Noise ratio used to be a measure of image quality. The PSNR between two images having 8 bits per pixel or sample in terms of decibels (dBs) is given by: PSNR = 10 log10

(9)

mean square error (MSE). Generally when PSNR is 40 dB or greater, then the original and the reconstructed images are virtually indistinguishable by human observers.

Figure 2: Capturing face in a video and add specifying the details

Figure 3: Perceiving Skin area

73

International Journal of Emerging Sciences 5(2), 68-75, June 2015

Figure 4: Detecting the face emotion using eyes and mouth

In all results presented, the per-class evaluation measure is exact, and the overall measure is accuracy.

5.

CONCLUSION

For video-based face recognition, foreground and background objects can be separated by detecting moving objects, which eliminating complex background facilitates tasks of face detection and recognition. In order to employ the advantages of DWT algorithm for multidimensional signals such as video, it is proper to use its structure only for the longest dimension. Then the procedure outputs the prepared coefficient and inputs the next sample to the buffer. The wavelet transform has been successfully applied to image and signal processing. The performances were obtained using different number of training images. The advantages of using PCA are that smaller representation of database because we only store the training images in the form of their projections, on the reduced basis. Noise is reduced because we choose the maximum variation basis and hence features like background with small variation are automatically ignored. This face identification system is based on easy to use, flexible and customizable face detection software. In terms of face recognition correctness, scalability and performance is impressive.

REFERENCES 1. Ruicong Zhi, Qiuqi Ruan, "A Comparative Study on Region-Based Moments for Facial Expression Recognition,", in Congress on Image and Signal Processing, Vol. 2, pp.600-604, 2008. 2. Irene Kotsia, Ioannis Pitas, “ Facial Expression Recognition in Image Sequences Using Geometric Deformation Features and Support Vector Machines” in IEEE Transactions on Image Processing 16(1): pp. 172-187, 2007. 3. Kakumanu.P., Nikolaos G. Bourbakis, “ A Local-Global Graph Approach for Facial Expression Recognition” . ICTAI, pp 685-692, 2006. 4. Aleksic. P.S., Aggelos K. Katsaggelos. “Automatic facial expression recognition using facial animation parameters and multistream HMMs”. IEEE Transactions on Information Forensics and Security 1(1): pp. 3-11, 2006 5. Spiros Ioannou, George Caridakis, Kostas Karpouzis, and Stefanos Kollias, “Robust Feature Detection for Facial Expression Recognition”, EURASIP Journal on Image and Video Processing, vol. 2007, Article ID 29081, 2007. 6. Pantic.M and L. J. M. Rothkrantz. “Automatic analysis of facial expressions: the state of the art,” IEEE Trans. on PAMI, vol. 22, no. 12, pp. 1424-1445, 2000. 7. A. Samal and P.A. Iyengar, “Automatic Recognition and Analysis of Human Faces and Facial Expressions: A Survey,” Pattern Recognition, vol. 25, no. 1, pp. 65-77, 1992.

74

D. Magdalene Delighta Angeline and I. Samuel Peter James

8. Tian.Y.L, T. Kanade, and J. F. Cohn. “Evaluation of Gabor-Wavelet-Based Facial Action Unit Recognition in Image Sequences ofIncreasing Complexity,” in Proc. of the Fifth IEEE Int. Conf. on AFGR, pp. 229-234, 2002. 9. Pantic.M and L. J. M. Rothkrantz. “Facial Action Recognition for Facial Expression Analysis From Static Face Images” in the proc. of IEEE Transactions on Systems, Man and Cybernetics- Part B: Cybernetics, Vol 34, No.3, 2004. 10. Ashish Kapoor, Yuan Qi and Rosalind W. Picard. “Fully Automatic Upper Facial Action Recognition” in Proceedings of the IEEE International Workshop on Analysis and Modeling of Faces and Gestures (AMFG’03), pp.195 – 202, 2003. 11. Xiaoyi Feng, Baohua Lv, Zhen Li, Jiling Zhang. “A Novel Feature ExtractionMethod for Facial Expression Recognition” Proceedings of the 2006 Joint Conference on Information Sciences, JCIS 2006, Kaohsiung, Taiwan, ROC, October 8-11, 2006. Atlantis Press 2006, ISBN 90-78677-01-5. 12. Shackleton M.A. and W. J. Welsh. “Classification of Facial Feature for Recognition”, in Proc. CVPR91, pp.573-579, 1991. 13. Terzopoulos.D. and K.Waters “Analysis and Synthesis of Facial Image Using Physical and Anatomical Models”, IEEE Transactions on PAMI, vol. l5, No.6, pp.569-579, 1993.

75