IJIGSP Vol. 9, No. 6, Jun. 2017
Cover page and Table of Contents: PDF (size: 292KB)
The electrocardiogram (ECG) will create the characteristic in the form of the wave’s peak pattern. The first peak and the next one in one ECG wave have their own value and names, namely PQRST peaks. The process of feature extraction is very significant to determine the certain pattern. The use of feature extraction will be useful to help to detect certain case, including the determination of PQRST peaks according to the ECG print-out. This study makes a method to determine the ECG peaks (PQRST), the heart rate, and ST-deviation according to the ECG graphic image. The input data is in the form of ECG graphic image which is derived from the ECG 12 lead record. This study employs segmentation method (grayscale and binary), morphology (dilation and erosion), and produce the graphic image which is read as the ECG signal in the pre-processing stage, and use the Pan-Tompkins algorithm for the feature extraction method. The result of the peak determination is validated by cardiologists. The validation shows that the result of up and down deflection computation from the isoelectric of each P, Q, R, S, and T wave has represented the ECG calculation clinically; including the calculation to determine the R-R interval, heart rate, and ST-deviation.[...] Read more.
In this paper, we propose a novel JPEG image Steganography algorithm based on partition schemes on image coefficient values. Our method selects the AC and DC coefficients of a JPEG image according to a channel selection method and then identifies appropriate coefficients to store the secret data-bits. As opposed to other reported works, in our algorithm each selected coefficient can store a variable number of data-bits that are decided using the concept called ‘Partition Scheme’. Experimental results indicate the suitability of the proposed algorithm as compared to other existing methods.[...] Read more.
In this paper, we propose a synthesis method for synthesizing the reconfigurable multiple patterns with the minimum number of antenna elements based on the state space model. The proposed method is to obtain the common element locations for the multiple patterns using fewer antenna elements within desired performance bounds. The proposed approach introduces the state-space method to represent the multiple patterns and then uses the multiple pattern data to construct a combined Hankel matrix which is used to estimate the model parameters from which the number of elements and the common element locations can be extracted. Numerical results show the effectiveness of the proposed methods.[...] Read more.
Due to rapid growth in image sizes, an alternate of numerically lossless coding named visually lossless coding is considered to reduce storage size and lower data transmission. In this paper, a lossy compression method on encrypted color image is introduced with undetectable quality loss and high compression ratio. The proposed method includes the Xinpeng Zhang lossy compression , Hierarchical Oriented Prediction (HOP), Uniform Quantization, Negative Sign Removal, Concatenation of 7-bit data and Huffman Compression. The encrypted image is divided into rigid and elastic parts. The Xinpeng Zhang elastic compression is applied on elastic part and HOP is applied on rigid part. This method is applied on different test cases and the results were evaluated. The experimental evidences suggest that, the proposed method has better coding performance than the existing encrypted image compressions, with 9.645 % reductions in bit rate and the eye perception is visually lossless.[...] Read more.
Multi-scale transform (MST) and sparse representation (SR) techniques are used in an image representation model. Image fusion is used especially in medical, military and remote sensing areas for high resolution vision. In this paper an image fusion technique based on shearlet transformation and sparse representation is proposed to overcome the natural defects of both MST and SR based methods. The proposed method is also used in different transformations and SR for comparison purposes. This research also investigate denoising techniques with additive white Gaussian noise into source images and perform threshold for de-noised into the proposed method. The image quality assessments for the fused image are used for the performance of proposed method and compared with others.[...] Read more.
Wireless Capsule Endoscopy (WCE) is one of the methods for examination of gastrointestinal (GI) disorders such as obscure GI bleeding, Crohns disease, polyps etc. WCE has been recognized as a less expensive and painless procedure for the diagnosis of GI tract. This paper examines the various image classifiers designed and developed for the purpose of endoscopy focusing specifically on WCE. It is revealed that designing a suitable image classifier is an important prerequisite for accurate and precise diagnosis of malignancy in WCE. The assessment on various image classifiers used for the diagnosis of pathologies in different parts of GI tract shows that classifiers have reduced the diagnosis time for medical experts and also provided reasonably accurate diagnosis of malignancy. However, correlating classifiers and related pathologies is still observed to be challenging. In view of the fact that early detection may decrease the mortality rate significantly, inclination towards computer aided diagnosis are expected to increase in future. There is a need for advanced research in the development of a robust computer aided diagnosis system, capable of diagnosis of various pathologies in GI tract with higher degree of accuracy and reliability. Further, the study depicts that a direct comparison of results of classifier such as accuracy, prediction, sensitivity, specificity and precision to evaluate its performance is challenging due to diversity of image databases. More research is needed to identify and reduce the uncertainties in the application of image classifier to improve the diagnosis accuracy.[...] Read more.
Development of advanced technology has increased the size of data and has also created different categories of data. Classifying these different categories of data is the need of the era. We have proposed a method of classifying the image database containing four categories of images like human face, airplane, cup and butterfly. Our approach involves steps like feature extraction, bag of feature creation, histogram representation and classification using decision tree. For feature extraction SIFT (Scale Invariant Feature Transform) algorithm is used since it is invariant to rotation, change of scale, illumination etc. After extracting the features the bag of features concept is used to group the features using k-means clustering algorithm. Then a histogram is plotted for each image in the image database which represents the distributions of data in different clusters. In the final step the most robust, simple and flexible decision tree algorithm is applied on the table created from the histogram plots to obtain the classification result. The experimental observations and the calculated accuracy proves that this method of classification works well for classifying an image dataset having different categories of images.[...] Read more.