Ajit S. Bopardikar

Cardiac arrhythmia detection using dynamic time warping of ECG beats in e-healthcare systems

Automatic real-time detection and classification of ECG patterns is of great importance in early diagnosis and treatment of life-threatening cardiac arrhythmia. In this paper, we have presented dynamic time warping (DTW) distance based approach for classification of arrhythmic ECG beats, with an aim of using it in smart-phone/mobile environment. The performance of the proposed method is tested on ECG beats of various arrhythmia types selected from MIT-BIH arrhythmia database. We have compared the proposed DTW approach using naïve Bayes classifier with relative band spectral power as feature. The DTW approach has shown superior performance compared to the naïve Bayes classifier. Furthermore, we have verified the performance of the DTW approach on down-sampled ECG beats in order to improve speed of the DTW algorithm. It is observed that the performance of the DTW approach did not deteriorate even after subsampling of ECG beats. The DTW with subsampling has been aimed at real-time arrhythmia detection in wearable mobile healthcare systems in telemedicine scenario for continuous monitoring of ECG records from cardiac patients.

No-Reference Compression of Genomic Data Stored in FASTQ Format

In this paper, we propose a system to compress Next Generation Sequencing (NGS) information stored in a FASTQ file. A FASTQ file contains text, DNA read and quality information for millions or billions of reads. The proposed system first parses the FASTQ file into its component fields. In a partial first pass it gathers statistics which are then used to choose a representation for each field that can give the best compression. Text data is further parsed into repeating and variable components and entropy coding is used to compress the latter. Similarly, Markov encoding and repeat finding based methods are used for DNA read compression. Finally, we propose several run length based methods to encode quality data choosing the method that gives the best performance for a given set of quality values. The compression system provides features for loss less and nearly loss less compression as well as compressing only read and read + quality data. We compare its performance to bzip2 text compression utility and an existing benchmark algorithm. We observe that the performance of the proposed system is superior to that of both the systems.

Detail warping based video super-resolution using image guides

Video scaling is a crucial video processing component that finds a variety of applications in video display, enhancement and analysis. Thus, video scaling has been an important research topic for several years now. With the emergence of hybrid-cameras that simultaneously capture a low-resolution (LR) video and periodic high-resolution (HR) stills, methods that use information in HR stills for enhancing the LR video are being studied. We present a multi-layer super-resolution (SR) method to rescale the input LR video with the help of a few HR still image guides. Resolution enhancement is achieved by blending the high frequency details into the input video frames that are scaled-up by a conventional, spatial scaling method. The missing high frequency details for each video frame are derived by spatio-temporal warping of the high frequency details extracted from HR still image guides. The Non-Local-Means (NLM) concept is used for computing the warping weights. The proposed SR method handles both global and local motions in a video, without requiring any explicit image registration. Experimental results comparing the proposed method with some of the existing video scaling techniques clearly points its strong potential.

Algorithm for DNA sequence compression based on prediction of mismatch bases and repeat location

For DNA sequence Compression, it has been observed that methods based on Markov modeling and repeats give best results. However, these methods tend to use uniform distribution assumption of mismatches for approximate repeats. We show that these replacements are not uniformly distributed and we can improve compression efficiency by using non uniform distribution for mismatches. We also propose a hash table based method to predict repeat location which works well for block based genomic sequence compression algorithms. The proposed methods give good compression gains. The method can be incorporated into any algorithm that uses approximate repeats to realize similar gains.

Selecting High Quality Frames for Super Resolution Reconstruction Using Perceptual Quality Metrics

Super-resolution involves the use of signal processing techniques to estimate the high-resolution (HR) version of a scene from multiple low-resolution (LR) observations. It follows that the quality of the reconstructed HR image would depend on the quality of the LR observations. The latter depends on multiple factors like the image acquisition process, encoding or compression and transmission. However, not all images are equally affected by a given type of impairment. A proper choice of the LR observations for reconstruction, should yield a better estimate of the HR image, over a naive method using all images. We propose a simple, model-free approach to improve the performance of super-resolution systems based on the use of perceptual quality metrics. Our approach does not require, or even assume, a specific realization of the SR system. Instead, we select the image subset with high perceptual quality from the available set of LR images. Finally, we present the logical extension of our approach to select the perceptually significant regions in a given LR image, for use in SR reconstruction.

Estimating cancer gene pathway proximity using network interaction

It is known that the gene level aberrations for a given cancer could vary across patients. As a result, a single therapy may not be suitable for every patient. However, these genetic aberrations may occur in similar pathways across patients. Therefore a study at pathway/subnetwork is more effective than at gene level. In this paper, we propose a method at this level to classify pathways (sub-networks) as functionally coupled and functionally independent. For this, we propose novel interaction measures. We show how these can be used to link and classify subnetworks using breast cancer as an example. Such methods will play an important role in patient stratification in order to develop personalized treatment options.

Identification of CpG islands in DNA sequences using supervised classification

In this paper, we propose a supervised classification approach based on Euclidean distance for identifying CpG islands in DNA sequences. We first extract features from the training data set which is extracted from annotated DNA sequences. CpG island locations in a test data sequence are identified by calculating Euclidean distance in the feature space. A moving window method has been used to screen the input test sequence. The performance of the proposed method is verified experimentally on EMBL human DNA database. Proposed approach gives superior performance results over most of the available CpG island detectors and has potential application in annotating CpG islands in large human sequences.

Alpha model based mixed pixel processing for view synthesis

The commercial success and acceptability of 3D technology will critically depend on the overall visual quality of the rendered images. Therefore, view synthesis techniques like Depth Image Based Rendering (DIBR) are crucial components of the 3D system chain. In this paper, we propose a fast α-model based approach that enables high quality DIBR. Given an image and depth map, a point based representation of the scene is obtained, where each point is associated with color, transparency and depth. A low complexity method estimates these attributes, making the system suitable for deployment in consumer electronics.

A method to identify dynamic zones for efficient control of HVAC systems

Energy Management System (EMS) in Smart Buildings is responsible for providing thermal comfort to its occupants in an energy efficient manner. To achieve this, the EMS usually divides the floor of a building into multiple zones with predefined static boundaries. Next, it derives the optimal Heating Ventilation and Air Conditioning (HVAC) control parameters based on the average environmental characteristics (e.g. average temperature, humidity and occupancy) for each zone. Because the HVAC acts based on the average conditions in a zone, static zones can display considerable variation within their boundaries from the desired ambient conditions. A more adaptive approach is therefore called for in order to reduce deviations from desired ambient conditions. In this paper, we propose a graph theoretic region growing approach to dynamically identify zones with similar ambient conditions. The boundary of these dynamic zones can be adapted continuously as the environmental conditions change. In this manner, the HVAC parameters can be adapted to zones with smaller variations, thereby achieving the desired ambient conditions across the floor with higher accuracy. The efficacy of the proposed approach is demonstrated for a single floor in a building using real time data collected from various sensors (humidity and temperature) and occupancy. Finally, dynamic zone information can be used to save energy. Simulation results indicate considerable energy savings.

Context based compression of FASTQ data

With advances in Next Generation Sequencing (NGS) technologies, the amount of genomic data produced is growing exponentially. Efficient compression is therefore vital for archiving, retrieval and transfer of raw sequencing data. NGS data consisting of sequence information along with the associated quality scores and sequence identifiers is stored in the FASTQ format. In this paper, we present lossless methdologies to compress components of a FASTQ file. The proposed system explores p rediction by partial matching methods (PPM) for building higher order context models to compress FASTQ data using adaptive arithmetic coding (AAC). We analyze crucial parameters in AAC to further improve overall compression of our system. We compare the performance of the proposed system with existing benchmarks for a sample data set of 6 standard FASTQ files. The proposed method provides gains of up to 15% compared to the best existing benchmark. We also show that the proposed methodology provides gains on overall storage space across the sample data set.