M. Pallikonda Rajasekaran

An unsupervised learning method with a clustering approach for tumor identification and tissue segmentation in magnetic resonance brain images

Novel SOM based FKM algorithm for tissue segmentation and tumor identification in magnetic resonance brain images (T1-w, T2-w, FLAIR and MPR sequences) is proposed through this work.Exact demarcation between tumor and edema region is characterized.Validation of the segmented results by an experienced radiologist.Cross comparison with FCM, SOM, FKM and other hybrid clustering algorithms using ten standard comparison parameters. Malignant and benign types of tumor infiltrated in human brain are diagnosed with the help of an MRI scanner. With the slice images obtained using an MRI scanner, certain image processing techniques are utilized to have a clear anatomy of brain tissues. One such image processing technique is hybrid self-organizing map (SOM) with fuzzy K means (FKM) algorithm, which offers successful identification of tumor and good segmentation of tissue regions present inside the tissues of brain. The proposed algorithm is efficient in terms of Jaccard Index, Dice Overlap Index (DOI), sensitivity, specificity, peak signal to noise ratio (PSNR), mean square error (MSE), computational time and memory requirement. The algorithm proposed through this paper has better data handling capacities and it also performs efficient processing upon the input magnetic resonance (MR) brain images. Automatic detection of tumor region in MR (magnetic resonance) brain images has a high impact in helping the radio surgeons assess the size of the tumor present inside the tissues of brain and it also supports in identifying the exact topographical location of tumor region. The proposed hybrid SOM-FKM algorithm assists the radio surgeon by providing an automated tissue segmentation and tumor identification, thus enhancing radio therapeutic procedures. The efficiency of the proposed technique is verified using the clinical images obtained from four patients, along with the images taken from Harvard Brain Repository.

Sensor grid applications in patient monitoring

Wireless Sensor Networks (WSNs) are finding an important role in patient monitoring in diverse environments including hospitals for post-operative patients and nursing homes for elderly patients. Sensor networking devices in WSNs are resource constrained since they have limited processing power and communication bandwidth. However, with a large number of such devices being deployed and aggregated over a wide area, WSNs have substantial data acquisition and processing capability. Thus, WSNs are important distributed computing resources that can be shared by different groups of patients in different environments. The emerging domain of WSNs with the grid extends the grid computing paradigm to the sharing of sensor resources in WSNs. In this perspective, by their very demand requirements and their socioeconomic impact, medical applications are certainly the most pertinent domain for using a wireless sensor grid. In this paper, we propose a wireless sensor grid architecture for monitoring the health status of different groups of patients to provide a platform for physicians and researchers to share information with distributed database and computational resources to facilitate analysis, diagnosis, prognosis and drug delivery.

An automated hybrid approach using clustering and nature inspired optimization technique for improved tumor and tissue segmentation in magnetic resonance brain images

In the domain of human brain image analysis, identification of tumor region and segmentation of tissue structures tend to be a challenging task. Automated segmentation of Magnetic Resonance (MR) brain images would be of great assistance to radiologist, as they minimize the complication evolved due to human interface and offer quicker segmentation results. Automated algorithms offer minimal time duration and lesser manual intervention to a radiologist during clinical diagnosis. Moreover, larger volumes of patient data could be assessed with the aid of an automated algorithm and one such algorithm is proposed through this research to identify the tumor region bounded between normal tissue regions and edema portions. The proposed algorithm offers a better support to a radiologist in the process of diagnosing the pathologies, since; it utilizes both optimization and clustering techniques. Bacteria Foraging Optimization (BFO) and Modified Fuzzy K − Means algorithm (MFKM) are the optimization and clustering techniques used to render efficient MR brain image analysis. The proposed combinational algorithm is compared with Particle Swarm Optimization based Fuzzy C − Means algorithm (PSO based FCM), Modified Fuzzy K − Means (MFKM) and conventional FCM algorithm. The suggested methodology is evaluated using the comparison parameters such as sensitivity, Specificity, Jaccard Tanimoto Co − efficient Index (TC) and Dice Overlap Index (DOI), computational time and memory requirement. The algorithm proposed through this paper has produced appreciable values of sensitivity and specificity, which are 97.14% and 93.94%, respectively. Finally, it is found that the proposed BFO based MFKM algorithm offers better MR brain image segmentation and provides extensive support to radiologists.

An IoT based patient monitoring system using raspberry Pi

In the recent development of, Internet of Things (IoT) makes all objects interconnected and it has been recognized as the next technical revolution. Some of the applications of Internet of Things are smart parking, smart home, smart city, smart environment, industrial places, agriculture fields and health monitoring process. One such application is in healthcare to monitor the patient health status Internet of Things makes medical equipments more efficient by allowing real time monitoring of patient health, in which sensor acquire data of patients and reduces the human error. In Internet of Things patients parameters get transmitted through medical devices via a gateway, where it is stored and analyzed. The significant challenges in the implementation of Internet of Things for healthcare applications is monitoring all patients from various places. Thus Internet o Things in the medical field brings out the solution for effective patient monitoring at reduced cost and also reduces the trade-off between patient outcome and disease management. In this paper discuss about, monitoring patients body temperature, respiration rate, heart beat and body movement using Raspberry Pi board.

Performance evaluation of radial basis function networks based on tree seed algorithm

This paper develops a Radial Basis Function Neural Network (RBFN) based on Tree Seed Optimization algorithm (TSA). The values of clustering centers, width and weights of the Radial Basis Function Neural Network are optimized by Tree Seed Algorithm. The proposed Radial Basis Function Neural Network optimization algorithm is tested on the application of numerical function approximation. The experimental result shows that the optimization of Radial Basis Function Neural Network with Tree Seed Algorithm has improved significance in attaining the faster convergence and also the extent of fitness has been improved.

An efficient hierarchical attribute set based encryption scheme with revocation for outsourcing personal health records in cloud computing

Enterprises will outsource their sensitive data in a cloud server due to the rapid development of cloud computing in the IT industry for last few years. It is attractive for the Personal Health Record (PHR) service providers to shift their PHR applications and storage into the cloud. Under encryption, it is excited to achieve fine-grained access control to PHR data in a scalable and efficient way. It also includes the problem of establishing access control for the encrypted data, and revoking or withdrawing the access rights from users when they are no longer authorized to access the encrypted data on cloud servers. But the use of one single Trusted Authority (TA) and Cipher text Policy (CP-ABE) are unable to manage multiple group owners for encryption process and access policy. In order to realize scalability, flexibility, and fine-grained access control of outsourced data in cloud computing, we leverage Hierarchical Attribute-Set-Based Encryption (HASBE). This HASBE extends the cipher text-policy Attribute-Set-Based Encryption (ASBE) with a hierarchical structure of users by means of compound attributes. The intended scheme achieves fine-grained, flexible and scalable data access control with the help of compound attributes of HASBE.

Application of adaptive neuro-fuzzy inference systems for MR image classification and tumour detection

In earlier days, Magnetic-Resonance (MR) brain image classification and tumour detection was done by humans. But, this classification is impractical for large amounts of data. The uses of intelligence techniques have shown great improvement. Hence, in this paper the ANFIS is applied for classification and detection. Decision making was performed in two stages: feature extraction using Principal Component Analysis (PCA) and ANFIS trained with the backpropagation gradient descent method in combination with the least-squares method. The performance of the ANFIS classifier is evaluated in terms of training performance and classification accuracies and the results confirms that the proposed ANFIS has potential in detecting the tumours.

Analysis of Particle Swarm Optimization Based 2D FIR Filter for Reduction of Additive and Multiplicative Noise in Images

Noise in digital images is the major cause of severe artifacts. Filter design for denoising applications can also be addressed with optimization techniques as conventional filters incur in this. Exploration and Exploitation capability features of the Meta Heuristic Optimization Techniques make them applicable to noise reduction in digital images. An increasing number of Meta Heuristic Optimization algorithms make it suitable for designing FIR filters. In the proposed method, Particle Swarm Optimization, a global optimizer algorithm was used in calculating the appropriate coefficients for 2D FIR Filter. The proposed filter was applied to standard test images for testing its noise suppression capability. Indicators of performance, such as Peak signal to noise ratio (PSNR) values and Structural Content (SC) were used in accessing the efficiency of the proposed method and to the adaptability of the method for removing different noise types. Thus a brief comparison for noise suppression in digital images with both multiplicative and additive noise types using PSO optimized 2D FIR filter is addressed in this paper.

Comparison of image compression techniques using huffman coding, DWT and fractal algorithm

Image compression is one of the advantageous techniques in different types of multi-media services. Image Compression technique have been emerged as one of the most important and successful applications in image analysis. In this paper the proposal of image compression using simple coding techniques called Huffman; Discrete Wavelet Transform (DWT) coding and fractal algorithm is done. These techniques are simple in implementation and utilize less memory. Huffman coding technique involves in reducing the redundant data in input images. DWT can be able to improve the quality of compressed image. Fractal algorithm involves encoding process and gives better compression ratio. By using the above algorithms the calculation of Peak signal to noise ratio (PSNR), Mean Square error (MSE) and compression ratio (CR) and Bits per pixel (BPP) of the compressed image by giving 512×512 input images and also the comparison of performance analysis of the parameters with that above algorithms is done. The result clearly explains that Fractal algorithm provides better Compression ratio (CR) and Peak Signal to noise ratio (PSNR).

Performance Evaluation of PET Image Reconstruction Using Radial Basis Function Networks

In this paper, for the reconstruction of the positron emission tomography (PET) images, Artificial Neural Network (ANN) method and Artificial Neural Network-Radial Basis Function (ANN-RBF) method are pursued. ANN is a dominant tool for demonstrating, exclusively when the essential data relationship is unfamiliar. ANN imitates the learning process of the human brain and can process problems involving nonlinear and complex data even if the data are imprecise and noisy. But, ANN calls for high processing time and its architecture needs to be emulated. So, ANN-RBF method is implemented which is a two-layer feed-forward network in which the hidden nodes implement a set of radial basis functions. Thus, the learning process is very fast. By the image quality parameter of peak signal-to-noise ratio (PSNR) value, the ANN method and the ANN-RBF method are compared and it was clinched that better results are obtained from ANN with RBF method.