A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Nasipuri, Mita
- 3D Face Recognition from Range Images Based on Curvature Analysis
Authors
1 Department of Computer Science and Engineering, Jadavpur University, IN
Source
ICTACT Journal on Image and Video Processing, Vol 4, No 3 (2014), Pagination: 748-753Abstract
In this paper, we present a novel approach for three-dimensional face recognition by extracting the curvature maps from range images. There are four types of curvature maps: Gaussian, Mean, Maximum and Minimum curvature maps. These curvature maps are used as a feature for 3D face recognition purpose. The dimension of these feature vectors is reduced using Singular Value Decomposition (SVD) technique. Now from calculated three components of SVD, the non-negative values of 'S' part of SVD is ranked and used as feature vector. In this proposed method, two pair-wise curvature computations are done. One is Mean, and Maximum curvature pair and another is Gaussian and Mean curvature pair. These are used to compare the result for better recognition rate. This automated 3D face recognition system is focused in different directions like, frontal pose with expression and illumination variation, frontal face along with registered face, only registered face and registered face from different pose orientation across X, Y and Z axes. 3D face images used for this research work are taken from FRAV3D database. The pose variation of 3D facial image is being registered to frontal pose by applying one to all registration technique then curvature mapping is applied on registered face images along with remaining frontal face images. For the classification and recognition purpose five layer feed-forward back propagation neural network classifiers is used, and the corresponding result is discussed in section 4.Keywords
Curvature Analysis, 3D Image, Image Registration, Face Recognition, FRAV3D Database.- A Novel Low Space Image Storing and Reconstruction Method by K-Means Clustering Algorithm
Authors
1 Department of Computer Science and Engineering, Government College of Engineering and Leather Technology, Kolkata, West Bengal, IN
2 Department of Computer Science and Engineering, Government College of Engineering and Textile Technology, Berhampore, West Bengal, IN
3 Department of Computer Science and Engineering, Jadavpur University, Kolkata-700032, West Bengal, IN
Source
International Journal of Technology, Vol 4, No 1 (2014), Pagination: 1-11Abstract
This paper presents a lossy image compression technique that proposes a novel approach for storing RGB color images which save 33% memory space compared to memory space requirement of conventional method of storing RGB images. The proposed method, first finds the most and least dominating color components among three Red, Green and Blue color channels for each RGB image and then for each pixel of the image, it finds the absolute difference between the most and least dominating color values and expresses the difference as a fraction of the most dominating color value of the pixel. All the fraction values are clustered into sixteen groups using K-Means Clustering algorithm and all centroids are stored as Header. The less significant two bits of each the other two color channels are modified according to the cluster information. These two modified color channels along with one Header are stored for each image. Thus 33% of the memory space requirement to store the original image could be saved using the proposed method. At the time of reconstruction of the image, according to the cluster information third color component is retrieved with the help of the header. The experimental result shows that the reconstructed images retain around 98.5±0.5% of the original image information. The method has been implemented using Matlab 7 and tested on one standard FRAV2D database and hundred natural images and this method could be applied to compress any RGB image.