172
Facial Expression Recognition Using ANN & Gabor Filter
Mr.Vaibhav Chaudhari ME Student, Computer Science
S.S.G.B.C.O.E.T Bhusaval, [email protected]
Prof. Y. S. Patil
Assistant Professor, Computer Science S.S.G.B.C.O.E.T Bhusaval,
[email protected]
Prof. D. D. Patil HOD Computer Science S.S.G.B.C.O.E.T Bhusaval,
[email protected]
ABSTRACT
Analysis and recognition of human facial expressions from images and video forms the basis for understanding image content at a higher semantic level. Expression recognition forms the core task of intelligent systems based on human–
computer interaction (HCI). The study explores the use of Artificial Neural Networks in performing expression recognition. The system analysis seven basic types of human expressions – neutral, happy, sad, disgust, anger, surprise and fear. The ability of humans to recognize a wide variety of facial expressions is unparalleled. Researchers in the recent past have been trying to automate this task on a computer, employing a combination of image/ video processing techniques, along with machine learning techniques like ANNs. The system has been made by doing changes in Gabor filter & ANN and result shows more accuracy to detect the Facial Expression..
Keywords
Gabor Filter, Face Detection, Face Recognition, Edge Detection, Lab Color Filter.
1. INTRODUCTION
Face is the primary focus of attention in social intelligence, playing a major role in conveying identity and emotions.
Although the ability to infer intelligence or character from facial appearance is suspect, the human ability to recognize face is remarkable. They had recognized thousands of faces learnt throughout our lifetime and identify familiar faces at a glance even after years of separation. This skill is quite robust, despite large changes in the visual stimulus due to viewing conditions, expression, aging and distractions such as glasses or changes in hair style or facial hair [1]. Computational models of face-recognition, in particular, are interesting because they can contribute not only to theoretical insights but also to practical applications. Computers that recognize faces could be applied to a wide variety of problems, including criminal identification, security systems, image and film processing, and human computer interaction. Unfortunately, developing a computational model of face recognition is quite difficult, because faces are complex, multidimensional and meaningful visual stimuli. The user should focus his attention
toward developing a sort of early, pre attentive pattern recognition capability that does not depend on having three- dimensional information or detailed geometry. He should develop a computational model of face recognition that is fast, reasonably simple, and accurate.
2. EXISTING SYSTEM
Fig.No.1. Existing System
The method was evaluated by using two widely used facial expression databases, i.e., Japanese Female Facial Expressions (JAFFE) and Cohn-Kanade (CKþ). We have used 10-fold cross validation to evaluate the performance of the proposed method. As discussed earlier, face detection was carried out on all images followed by scaling to bring the face to a common resolution. Facial landmarks were detected and salient facial patches were extracted from each face image.
During training stage, a SVM classifier was trained between each pair of expressions. Here the training data were the concatenated LBP histogram features extracted from the salient patches containing discriminative characteristics between the given pair of expression classes. Similarly, 6C2 numbers of SVM classifiers were constructed and used for evaluating the performance on the test-set The Cohn-Kanade
173
database contains both male and female facial expression image sequences for the six basic emotions. In our experiments, the last image from each sequence was selected where the expression is at its peak intensity. The number of instances for each expression varies according to its availability. In our experiments on CKþ database, we used 329 images in total: anger (41), disgust (45), fear (53), happiness (69), sadness (56), and surprise (65).
Fig.No.2 Representation of Accuracy Table 1:- Recognition Rate for Existing System
Facial Expression Recognition Accuracy (%)
Happy 70
Disgust ++- 75
Anger 70
Sad 45
Neutral 65
Average 65
3. PROPOSED SYSTEM
The aim of this project is to recognize the face expression of the test image with the help of artificial neural network (ANN). For the face feature extraction and expression recognition step, it used the Gabor filter method, which utilized a set of filter bank.
Fig.No. 3. Basic Structure of System
Artificial neural networks (ANNs) or connectionist systems are computing systems inspired by the biological neural networks that constitute animal brains. Such systems learn (progressively improve performance) to do tasks by considering examples, generally without task-specific programming. For example, in image recognition, they might learn to identify images that contain cats by analyzing example images that have been tagged "cat" or "no cat" and using the analytic results to identify cats in untagged images.
They have found most use in applications difficult to express in a traditional computer algorithm using ordinary rule-based programming. An ANN is based on a collection of connected units called artificial neurons, (analogous to axons in a biological brain). Each connection (synapse) between neurons can transmit an unidirectional signal with an activating strength that varies with the strength of the connection. If the combined incoming signals (from potentially many transmitting neurons) are strong enough, the receiving (postsynaptic) neuron activates and propagates a signal of to downstream neurons connected to it. Typically, neurons are organized in layers, and signals travel from the first (input), to the last (output) layer. Over time, the goal shifted to matching specific mental abilities, leading to specific deviations from biology, such as back propagation. Neural networks have been used on a variety of tasks, including computer vision, speech recognition, machine translation, social network filtering, playing board and video games, medical diagnosis and many other domains. In addition to receiving and sending signals, units may have state, generally represented by real numbers, typically between 0 and 1. A threshold or limiting function may govern each connection and neuron, such that the signal must exceed the limit before propagating.
174
Fig.No. 4. Artificial Neural Network Model In image processing, a Gabor filter, named after Dennis Gabor, is a linear filter used for texture analysis, which means that it basically analyses whether there are any specific frequency content in the image in specific directions in a localized region around the point or region of analysis.
Frequency and orientation representations of Gabor filters are similar to those of the human visual system, and they have been found to be particularly appropriate for texture representation and discrimination. In the spatial domain, a 2D Gabor filter is a Gaussian kernel function modulated by a sinusoidal plane wave. Simple cells in the visual cortex of mammalian brains can be modeled by Gabor functions. Thus, image analysis with Gabor filters is thought to be similar to perception in the human visual system. Since the main purpose of this project is facial expression recognition , therefore, the sample pictures are taken under special consideration to ease up the face detection process. Each picture is taken under the condition that, only face is the largest skin colored continuous object in the frame. There are two sets of pictures. One is used for training purpose and another is used for testing. The pictures are classified in the following expressional classes.
1. Image01 to Image13 = Happy
2. Image14 to Image24= Disgust 3. Image25 to Image34= Anger 4. Image35 to Image43 = Sad 5. Image44 to Image50 = Neutral
Table 2: Recognition Rate for various expressions of proposed method
Facial Expression Recognition Rate (%)
Happy 92.30
Disgust 80.71
Anger 83.20
Sad 76.73
Neutral 72.00
Average 80.98
Happy Recognition
Anger Recognition
175
Disgust Recognition
Normal Recognition
Sad Recognition
4. CONCLUSION
From our experimental study the facial expression recognition system worked well while the training image and the to-be- recognized image were the face of the same person in good lighting condition. However, it also demonstrated its limitation. Due to it depended on the gray level similarity of the training image set and the given test image, so it would be influenced greatly while the light, angle, face size and face
color were changed in either training database or the captured test image. The proposed method was tested on database of 10 different persons with different expressions. The proposed ANN and gabor filter based method has the greater accuracy with consistency. The recognition rate was greater even with the small number of training images which demonstrated that it is fast, relatively simple, and works well in a constrained environment.
ACKNOWLEDGMENT
First and foremost I would like to acknowledge with due courtesy the various sources consulted in the preparation of this work. I take immense pleasure in thanking Principal Dr.
R. P. Singh for his constant encouragement and support for carrying out this work. I wish to express my deep sense of gratitude to Prof. Yogesh. S. Patil for his able guidance, cordial support, valuable information and useful suggestions, which helped me in completing this project work in time. I am obliged to staff members of the Computer Science and Engineering Department for the valuable assistance provided by them. I am also grateful for their cooperation during the period of my project work. Finally, yet importantly, I would like to express my heartfelt thanks to my beloved parents for their blessings and constant encouragement without which this work would not be possible. I will be highly obliged for any positive criticism and suggestions from the readers regarding usefulness of this project work.
REFERENCES
[1] P. N. Belhumeur, J. P. Hespanha, and D. J.
Kriegman, “Eigenfaces versus Fisherfaces:recognition using class specific linear projection,” IEEE Trans. Pattern Anal. Mach.
Intell., vol. 19, no. 7, pp. 711–720, Jul. 1997.
[2] Md. FirdausHashim, Md. Rizon and putehsaad.
“Face recognition using Eigenfaces and neural networks”. American journal of applied sciences, November 2006.
[3] W.S. Yambor, B.A. Draper and J.R. Beveridge,
“Analyzing PCA-based face recognition algorithms:
eigenvector selections and distance measures”, in proc. Second Workshop Empirical Evaluation Computer Vision, 2000.
[4] Z. Sun, G.Bebis, X. Yuan and S.J. Louis “genetic feature subset selection for gender classification: A comparison study”, in Proc. Sixth IEEE Workshop Applications Computer Vision, 2002.
[5] A. Pentland, T. Starner, N. Etcoff, N. Masoiu, O.
Oliyide and M. Turk, “Experiments with eigenfaces”, in proc. Looking at People Workshop Int. Joint Artificial intelligence, 1993.
[6] D.L. Swets and J.J. Weng, “Using Discriminant Eigen features for image retrieval”, IEEE Trans.
Pattern Anal.Machine Intel, vol. 18, PP. 831-836, 1996.
176
[7] A.G. Ramakrishnan, S. Kumar Raja and H.V.
Raghu Ram, “Neural network-based segmentation of textures using Gabor features,” Proc. 12th IEEE Workshop on Neural Networks for Signal Processing, PP. 365 - 374, 2002.
[8] M.Turk, W. Zhao, R.chellappa, “Eigenfaces and Beyond”, Academic press, 2005.
[9] B. Moghaddam, “Principal manifolds and probabilistic subspaces for visual
[10] recognition”, IEEE transaction on pattern Anal.
Machine intelligence, vol.24, no.6, PP. 780-788, 2002.
[11] O. Deniz, M. Castrillfion and M. Hernfiandez,
“Face recognition using independent component analysis and support vector machines”, pattern recognition letters, vol.24, PP. 2153-2157, 2003.
[12] A Survey”, IEEE Transaction on Pattern analysis and machine Intelligence, vol.24, No.1, PP.24- 58,2002.
About Authors
Mr. Vaibhav Chaudhari, he received his BE from. SSGB College of Engineering and currently perusing. ME from Computer Engineering Department, SSGB College of Engineering Bhusaval India.. His main area of interest is Software Development and Artificial Neural Network.
Mr. Y.S.Patil Received his ME and currently working as Assistant Professor Computer Engineering Department, SSGB College of Engineering, Bhusaval.
Mr. D.D.Patil Received his M. Tech and currently working as Head of Computer Engineering Department, SSGB College of Engineering, Bhusaval.
