Mutual Correlation-Based Anonymization for Privacy Preserving Medical Data Publishing

Currently the industry is focused on managing, retrieving, and securing massive amounts of data. Hence, privacy preservation is a significant concern for those organizations that publish/share personal data for vernacular analysis. In this chapter, the authors presented an innovative approach that makes use of information gain of the quasi attributes with respect to sensitive attributes for anonymizing the data, which gives the fruitfulness of an attribute in classifying the data elements, which is a two-way correlation among attributes. The authors show that the proposed approach preserves better data utility and has lesser complexity than former methods.

A Novel Approach for Computer-Aided Diagnosis for Distinction Between Benign and Malignant of Lung Nodules Based on Machine Learning Techniques

A new method is proposed to classify the lung nodules as benign and malignant. The method is based on analysis of lung nodule shape, contour, and texture for better classification. The data set consists of 39 lung nodules of 39 patients which contain 19 benign and 20 malignant nodules. Lung regions are segmented based on morphological operators and lung nodules are detected based on shape and area features. The proposed algorithm was tested on LIDC (lung image database consortium) datasets and the results were found to be satisfactory. The performance of the method for distinction between benign and malignant was evaluated by the use of receiver operating characteristic (ROC) analysis. The method achieved area under the ROC curve was 0.903 which reduces the false positive rate.

Intrusion Detection System in Wireless Sensor Networks for Wormhole Attack Using Trust-Based System

Intrusion detection in wireless sensor network (WSN) has been a critical issue for the stable functioning of the networks during last decade. Wireless sensors are small and cheap devices that have a capacity to sense actions, data movement, and communicate with each other. It is a self-governing network that consists of sensor nodes deployed in a particular environment, which has wide applications in various areas such as data gathering, military surveillance, transportation, medical system, agriculture, smart building, satellite communication, and healthcare. Wormhole attack is one of the serious attacks, which is smoothly resolved in networks but difficult to observe. There are various techniques used to detect the malicious node such as LITEWORP, SAM, DelPHI, GRPW, and WRHT. This chapter focuses on detection methods for wormhole attacks using trust-based systems in WSN.

A Hybrid Watermarking Technique for Copyright Protection of Medical Signals in Teleradiology

Today, an individual's health is being monitored for diagnosis and treatment of diseases upon analyzing various medical data such as images and signals. Modifications of this medical data when it is transferred over an open communication channel or network leads to deviations in diagnosis and creates a serious health issue for any individual. Digital watermarking techniques are one of the solutions for providing protection to multimedia contents. This chapter gives requirements and various techniques for the security of medical data using watermarking. This chapter also demonstrates a novel hybrid watermarking technique based on fast discrete curvelet transform (FDCuT), redundant discrete wavelet transform (RDWT), and discrete cosine transform (DCT). This watermarking technique can be used for securing medical various types of medical images and ECG signals over an open communication channel.

Wireless Enhanced Security Based on Speech Recognition

Security is the most notable fact of all computerized control gadgets. In this chapter, a voice ID computerized gadget is utilized for the security motivation using speech recognition. Mostly, the voices are trained by extracting mel frequency cepstral coefficient feature (MFCC), but it is very sensitive to noise interference and degrades the performance; hence, dynamic MFCC is used for speech and speaker recognition. The registered voices are stored in a database. When the device senses any voice, it cross checks with the registered voice. If any mismatches occur, it gives an alert to the authorized person through global system for mobile communication (GSM) to intimate the unauthorized access. GSM works at a rate of 168 Kb/s up to 40 km and it operates at different operating frequencies like 800MHz, 900MHz, etc. This proposed work is more advantageous for the security systems to trap the unauthorized persons through an efficient communication.

Early Detection of Breast Cancer Using Image Processing Techniques

Breast cancer is the second leading cause of death among women according to Cancer Facts and Figures. In order to increase the survival rate of women due to breast cancer, early and accurate detection of breast cancer is very essential. The quality of the image acquired through different breast imaging modalities: mammography, ultrasonography, and magnetic resonance imaging (MRI) have drawbacks which reduces the efficiency of accurate detection. As per the literature survey, the quality of the image acquired through the breast imaging modalities is not optimal for the accurate detection of cancer at the early stage. Digital image enhancement methods have been widely used in radiology in order to enhance the image quality. Considerable research has been undertaken in the development of enhancement of the image to assist radiologists in the identification of breast abnormalities. In order to further improve the efficiency of detection, diagnosis, and treatment of cancer, the quality of the image has to be improved.

Secure Access to Biomedical Images

This chapter introduces a framework for secure access to biomedical images. Biomedical images are acquired using a vast array of imaging techniques depending upon the specific application. A magnetic resonance spatial domain image is acquired by taking inverse weighted Fourier transform of raw frequency domain data generated by the modality. After correction, these images are stored in a standard format. The access to these stored images is typically subjected to authorization. Medical information in biomedical images needs to be protected in both stored form and in transmission. Encryption technologies are used to secure information whereas compression technologies are used to reduce the information without affecting the contents. In this chapter, a cryptocompression system is proposed which integrates both encryption and compression to fulfill the requirements of electronic protected health information records.

Security Risks of Biomedical Data Processing in Cloud Computing Environment

The emergence of new innovations in technology changes the rate of data generated in health-related institutions and the way data should be handled. As such, the amount of data generated is always on the increase, which demands the need of advanced, automated management systems and storage platforms for handling large biomedical data. Cloud computing has emerged as the promising technology for present and future that can handle large amount of data and enhance processing and management of the data remotely. One of the disturbance concerns of the technology is the security of the data. Data in the cloud is subject to security threats, and this has highlighted the need for exploring security measures against the threats. The chapter provides detailed analysis of cloud computing deployment strategies and risks associated with the technology and tips for biomedical data storage and processing through cloud computing services.

Filtration and Classification of ECG Signals

Electrocardiogram (ECG) is a kind of process of recording the electrical activity/signals of the heart with respect to the time. ECG conveys a wide amount of information related to the structure and functions of the heart, its electrical conduction processes. ECG is a diagnostic tool that the doctors and medical professionals use to measure patients' heart activity by paying attention to the electric current flowing in the heart. Due to the presence of noises, it needs to carry out the filtration process. Filtration is the process of keeping the components of the signals of desired frequencies by setting up an “fc” value and removing the components apart from the said “fc” frequency. It is required to eliminate the noise level from the ECG signal, such that the resultant ECG signal must be free from noises. All these techniques and algorithms have their advantages and limitations which are discussed in this chapter.

Automatic Detection of Tumor and Bleed in Magnetic Resonance Brain Images

Brain tumor and intracerebral hemorrhage are major causes for death among the people. Brain tumor is the growth of abnormal cells multiplied in an uncontrolled manner in brain. Magnetic resonance imaging (MRI) technique plays a major role for analysis, diagnosis, and treatment planning of abnormalities in the brain. Bleed is detected manually by radiologists, but it is laborious, time-consuming, and error prone. The automatic detection method was performed to detect the tumor as well as bleed in brain under a single system. The proposed method includes image acquisition, pre-processing, patch extraction, feature extraction, convolutional neural network (CNN) classification, and fuzzy inference system (FIS) to detect the abnormality with reduced classification loss percentage. This chapter is compared with the existing system of tumor detection using convolution neural network based on certain features such as skewness, kurtosis, homogeneity, smoothness, and correlation.