|Title: Enhancing Passive Stereo Face Recognition using PCA and Fuzzy C-Means Clustering|
|Author(s): Lim Eng Aik, Tan Wee Choon|
|Pages: 1-5||Paper ID: 112404-5858-IJVIPNS-IJENS||Published: August, 2011|
Abstract: This paper presents a passive stereo vision face recognition system which uses stereo camera to detect and recognize a person’s face. The propose algorithm improves classical monocular 2D face recognition techniques by additionally considering the facial 3D surface, which is rather stable under various illumination conditions. Initially, individual faces are detected and their facial 3D surfaces are reconstructed from the stereo images. Next, the 3D face is then composed into its 2D image data with appropriate depth data and then decomposed into its principle components. The principle components are used to recognize a 3D face by comparing characteristics of the current face to those faces available in database. The results of our approach show a good improvement in recognition rate. For evaluation purposes, we then compared the performance of our approach to a classical face recognition algorithm and observed that the recognition rate increased on average by 9.03 percent.
|Keywords: Stereo vision, fuzzy clustering, principle component analysis, face recognition.|
|Full Text (.pdf) | 584 KB|
|Title: MRI Brain Images Segmentation Based on Optimized Fuzzy Logic and Spatial Information|
|Author(s): Indah Soesanti, Adhi Susanto, Thomas Sri Widodo, Maesadji Tjokronagoro|
|Pages: 6-11||Paper ID: 115804-7272-IJVIPNS-IJENS||Published: August, 2011|
Abstract: In this paper, a modified fuzzy c-means (FCM) clustering for medical image segmentation is presented. A conventional FCM algorithm does not fully utilize the spatial information in the image. In this research, we use a FCM algorithm that incorporates spatial information into the membership function for clustering. The spatial function is the summation of the membership function in the neighborhood of each pixel under consideration. The advantages of the method are that it is less sensitive to noise than other techniques, and it yields regions more homogeneous than those of other methods. This technique is a powerful method for noisy image segmentation. Originality of this research is the methods applied on a normal MRI brain image and a glioma MRI brain images, and analyze the area of tumor from segmented images. The results show that the method effectively segmented Magnetic Resonance Imaging (MRI) brain images with spatial information, and the segmented normal and glioma MRI brain images can be analyzed for diagnosis purpose. And, the area of abnormal mass is identified from 9.65 to 27.71 cm2.
|Keywords: Adaptive image segmentation, fuzzy logic, FCM clustering, MRI brain image, spatial information.|
|Full Text (.pdf) | 464 KB|
|Title: Neural Network based Clustering using Visual Features of Characters’ Shape in Image|
|Author(s): Safdar Zaman, Wolfgang Slany, S. Nadeem Ahsan, Farhan Hyder, Farukh Nadeem|
|Pages: 12-20||Paper ID: 118704-5656-IJVIPNS-IJENS||Published: August, 2011|
Abstract: Clustering gathers similar objects. A Character can also be treated as object and can be recognized in the image through its visual features. In this work, characters of the Urdu script are clustered on the basis of 18 different visual features. A Kohonen Self Organizing Map is used for clustering with four different topologies of sizes 6x5, 8x7, 9x8, and 10x10. Each topology is checked under 75, 100, 150 and 200 numbers of epochs. 30 Urdu characters make 106 different shapes due to the four different positions in the word. These 106 shapes are then classified into 53 general classes based on graphical similarity. The shape of each class comprises features for its description. Considering only 18 features of each shape, 53 general classes are then grouped into clusters using a Kohonen Self Organizing Map (K-SOM). The above mentioned work has been implemented in MATLAB.
|Keywords: Character’s shape, Features, Clustering, Kohonen-SOM, Topology.|
|Full Text (.pdf) | 1,507 KB|