Chikkanan Eswaran

Data Encryption Using Event-related Brain Signals

A method based on event-related brain signal is used for data encryption. The idea is to shuffle the Huffman tree using an encryption key generated by electroencephalogram (EEG) signals recorded when the user perceives a common black and white line picture. As different persons have different thought processes, the generated key is unique to each individual and hence the encryption is robust to fraudulent attacks as compared to other encryption systems. Further, as Huffman tree is used to encode the data during encryption, the method achieves both compression and encryption. This pilot study has shown the huge potential of the method as it is impossible to be compromised.

Performance Optimization of Video Coding Process on Multi-Core Platform Using Gop Level Parallelism

High definition video applications often require heavy computation, high bandwidth and high memory requirements which make their real-time implementation difficult. Multi-core architecture with parallelism provides new solutions to implementing complex multimedia applications in real-time. It is well-known that the speed of the H.264 encoder can be increased on a multi-core architecture using the parallelism concept. Most of the parallelization methods proposed earlier for these purposes suffer from the drawbacks of limited scalability and data dependency. In this paper, we present a result obtained using data-level parallelism at the Group-Of-Pictures (GOP) level for the video encoder. The proposed technique involves each GOP being encoded independently and implemented on JM 18.0 using advanced data structures and OpenMP programming techniques. The performance of the parallelized video encoder is evaluated for various resolutions based on the parameters such as encoding speed, bit rate, memory requirements and PSNR. The results show that with GOP level parallelism, very high speed up values can be achieved without much degradation in the video quality.

Improving Simplified Fuzzy ARTMAP Performance Using Genetic Algorithm for Brain Fingerprint Classification

A genetic algorithm is proposed for ordering the input patterns during training for simplified fuzzy ARTMAP (SFA) classifier to improve the individual identification classification performance using brain fingerprints. The results indicate improved classification performance as compared to the existing methods for pattern ordering, namely voting strategy and min-max. As the ordering method is general, it could be used with any dataset to obtain improved classification performance when SFA is used.

Gait recognition using occluded data

Gait is an attractive biometrics for use in monitoring and surveillance applications. In such settings, occlusion is common and may affect recognition. This paper investigates the performance of gait using occluded data. To reconstruct the data, interpolation is applied to the occluded data using the Support Vector Machines for Regression (SVR) framework. Then the Principal Component Analysis (PCA) and Canonical Analysis (CA) are applied to reduce the dimensionality of the reconstructed data and classification. Comparison is made between the recognition accuracy rates obtained using the occluded and visible data of the same subject.

Performance Comparison of Various Garbage Collectors on JVM for Web Services

The current demand of e-commerce has increased the requirement of efficiency on Java server-side applications. Hence, the continuous availability and the good response-time of the Java virtual machine are needed to satisfy the continuous incoming request from remote clients. This project studied the optimization of garbage collector in Java virtual machine particularly on 3-tier Java server-side application to increase the throughput of real times processing, full utilization of CPU time and more memory efficiency to handle more workloads. The characteristics and the architectures of five JikesRVM garbage collectors were studied. They are CopyMS, GenMS, SemiSpace, GenCopy and MarkSweep. The best performing garbage collectors was determined and the main causes of their overheads were identified. The benchmarking suite, SPECjbb2000 was used to emulate a 3-tier Java server-side application. The performance of the five different garbage collectors on SPECjbb2000 for large and small memory size was compared. In conclusion, CopyMS is found to have the best average throughput for large heap size whereas, GenMS has the best overall performance in memory constraint and can handle the most workloads. On the other hand, GenCopy and SemiSpace demonstrated higher efficiency in handling light workloads. The performance of garbage collectors is proportional to the heap size used. Memory fragmentation and long pause time are two main challenges to be overcome for increasing the application performance. Future works for optimizing garbage collectors are recommended at the end of this report

Distinguishing Twins by Gait via Jackknife-Like Validation in Classification Analysis

This paper is about analysing the uniqueness of twins by gait biometric. The motivation arises due to twins, having facial similarity may lay difficulties to a video-based recognition system employing face biometric. Gait, a biometric based on the way a person walk, can perhaps be a useful descriptor. Due to the small size data set, classification via leave-one-out cross validation may not be sufficient to test gait’s viability as a descriptor for twins. Thus, this paper proposes a jackknife-like validation in a matched-pair classification. Comparing between the results of both validation approaches, results of the proposed method have shown to be promising. The results perhaps may point to the uniqueness of each individual twin by gait biometric.

Optimization of JVM by Dynamic Thread Prioritization for Web Services

The explosive growth of e-business activities implemented over Web services has created a need for optimizing the throughput of Web server applications. Due to its portability and multithreading capability, Java has become the popular language for developing Web applications. The performance of Java application greatly depends on its bytecode interpreter namely Java virtual machine (JVM). Optimization has been studied on several areas such as thread synchronization, thread scheduler and garbage collections. This paper presents outcome of an implementation of priority mechanism that suits JikesRVM. The main objective is to obtain a simulated throughput improve on SPECjbb2000 benchmarking suite. The result of this project reveals an improvement of 3%-4% on throughput after the implementation of the new priority mechanism. Insufficient thread information and additional overhead are the two main challenges to overcome for this mechanism to achieve higher efficiency. From the result obtained, errors such as deadlock, starvation and priority inversion may occur if system thread priority level is assigned to be lower than the application thread. To further utilize the priority mechanism in JikesRVM suggested by this project, prioritization policy switching, multi-parameters reference and implementation of algorithm in other thread queues are possible

Matched-Pair Classification Analysis on Siblings' Gait Biometric Data

This paper presents a supervised classification task on gait biometric of siblings’ data sets. This task, which we refer to as matched-pair classification, evaluates the within pair differences in terms of the data set via jackknifing. A misclassification rate (MCR), which measures the percentage of misclassification of one sib compared to the other, gives an estimate on the potential uniqueness of gait for a person, particularly in twins. By this approach, the MCR values are mostly in the range of 90% for a data set of twins and in the range of 80% for a data set of non-twin siblings. When compared to the standard Leave-One-Out (LOO) classification, the MCR values of the proposed approach are higher than the LOO classification, which may suggest its potential use in machine learning with regard to biometric-based systems.

Gait classification of twins and non-twins siblings

This paper presents a classification analysis of gait biometric on twins and non-twins siblings. The aim of this paper is to investigate the existence or inexistence of similarity in the gait of twins and compare it to the gait of non-twins siblings. The motivation behind this paper is that a video-based surveillance system may not be able to rely on face biometric alone when dealing with twins. The features used are the angular displacement walking trajectories of lower limbs. Also this paper proposes a gait cycle normalization task via Bezier polynomial root-finding and re-sampling to ensure a robust analysis against differences in walking speed. Two established classifiers, the linear discriminant analysis (LDA) and k-nearest neighbor are used to classify the data sets of twins and non-twins siblings. Results may indicate that there is similarity in the gait of twins.

Recognizing Individual Sib in the Case of Siblings with Gait Biometric

Gait is another potential human biometrics to look into whenever face recognition fails in video-based systems as is the case with siblings that have similar faces. We perform analyses on 10 pairs of siblings where their faces are assumed to have similarities. Our gait features are the angular displacement trajectories of walking individuals. We apply smoothing with the Bezier polynomial in our root-finding algorithm for accurate gait cycle extraction. Then, we apply classification using two different classifiers; the linear discriminant analysis (LDA) and the k-nearest neighbour (kNN). The best average correct classification rate (CCR) is 100% with a city-block distance kNN classifier. Hence, it is suggested that in the case where face recognition fails, gait may be the better alternative for biometric identification.