Martin Glavin

Rear-Lamp Vehicle Detection and Tracking in Low-Exposure Color Video for Night Conditions

Automated detection of vehicles in front is an integral component of many advanced driver-assistance systems (ADAS), such as collision mitigation, automatic cruise control (ACC), and automatic headlamp dimming. We present a novel image processing system to detect and track vehicle rear-lamp pairs in forward-facing color video. A standard low-cost camera with a complementary metal-oxide semiconductor (CMOS) sensor and Bayer red-green-blue (RGB) color filter is used and could be utilized for full-color image display or other color image processing applications. The appearance of rear lamps in video and imagery can dramatically change, depending on camera hardware; therefore, we suggest a camera-configuration process that optimizes the appearance of rear lamps for segmentation. Rear-facing lamps are segmented from low-exposure forward-facing color video using a red-color threshold. Unlike previous work in the area, which uses subjective color threshold boundaries, our color threshold is directly derived from automotive regulations and adapted for real-world conditions in the hue-saturation-value (HSV) color space. Lamps are paired using color cross-correlation symmetry analysis and tracked using Kalman filtering. A tracking-based detection stage is introduced to improve robustness and to deal with distortions caused by other light sources and perspective distortion, which are common in automotive environments. Results that demonstrate the systems high detection rates, operating distance, and robustness to different lighting conditions and road environments are presented.

Echocardiographic speckle reduction comparison

In this paper, a detailed description and comparison of speckle reduction of medical ultrasound, and in particular echocardiography, is presented. Fifteen speckle reduction filters are described in a detailed fashion to facilitate implementation for research and evaluation. The filtering techniques considered include anisotropic diffusion, wavelet denoising, and local statistics. Common nomenclature and notation are adopted, to expedite comparison between approaches. Comparison of the filters is based on their application to simulated images, clinical videos, and a computational requirement analysis. The ultrasound simulation method provides a realistic model of the image acquisition process, and permits the use of a noise-free reference image for comparison. Application of objective quality metrics quantifies the preservation of image edges, overall image distortion, and improvement in image contrast. The computational analysis quantifies the number of operations required for each speckle reduction method. A speed-accuracy analysis of discretization methods for anisotropic diffusion is included. It is concluded that the optimal method is the OSRAD diffusion filter. This method is capable of strong speckle suppression, increasing the average SNRA of the simulated images by a factor of two. This method also shows favorable edge preservation and contrast improvement, and may be efficiently implemented.

Intra-Vehicle Networks: A Review

Automotive electronics is a rapidly expanding area with an increasing number of safety, driver assistance, and infotainment devices becoming standard in new vehicles. Current vehicles generally employ a number of different networking protocols to integrate these systems into the vehicle. The introduction of large numbers of sensors to provide driver assistance applications and the associated high-bandwidth requirements of these sensors have accelerated the demand for faster and more flexible network communication technologies within the vehicle. This paper presents a comprehensive overview of current research on advanced intra-vehicle networks and identifies outstanding research questions for the future.

Quasi-Multistatic MIST Beamforming for the Early Detection of Breast Cancer

Microwave imaging via space-time (MIST) beamforming has been shown to be one of the most promising imaging modalities for detecting small malignant breast tumors. This paper outlines two modifications to the MIST system developed by Hagness for the early detection of breast cancer, resulting in a quasi-multistatic MIST beamformer (multi-MIST). Multistatic MIST beamforming involves illuminating the breast with an ultrawideband (UWB) signal from one antenna while collecting the reflections at an array of antennas, as opposed to traditional monostatic MIST beamforming where only the transmitting antenna records the reflections from the breast. In order to process the multistatic data, traditional data-adaptive artifact removal algorithms have to be modified to accommodate signals from all antennas. Also, the MIST beamforming algorithm, which spatially focuses the signal and compensates for frequency-dependent propagation effects, has to be modified. The algorithms are tested on a 2-D anatomically accurate finite-difference time-domain model of the breast. The multi-MIST beamformer described here is shown to offer an improved signal to clutter ratio when compared to the traditional monostatic MIST beamformer.

Accuracy of fish-eye lens models

The majority of computer vision applications assumes that the camera adheres to the pinhole camera model. However, most optical systems will introduce undesirable effects. By far, the most evident of these effects is radial lensing, which is particularly noticeable in fish-eye camera systems, where the effect is relatively extreme. Several authors have developed models of fish-eye lenses that can be used to describe the fish-eye displacement. Our aim is to evaluate the accuracy of several of these models. Thus, we present a method by which the lens curve of a fish-eye camera can be extracted using well-founded assumptions and perspective methods. Several of the models from the literature are examined against this empirically derived curve.

Data Independent Radar Beamforming Algorithms for Breast Cancer Detection

Ultra wideband (UWB) Microwave imaging is one of the most promising emerging imaging technologies for breast cancer detection, and is based on the dielectric contrast between normal and cancerous tissues at microwave frequencies. UWB radar imaging involves illuminating the breast with a microwave pulse and re∞ected signals are used to determine the presence and location of signiflcant dielectric scatterers, which may be representative of cancerous tissue within the breast. Beamformers are used to spatially focus the re∞ected signals and to compensate for path dependent attenuation and phase efiects. While these beamforming algorithms have often been evaluated in isolation, variations in experimental conditions and metrics prompts the assessment of the beamformers on common anatomically and dielectrically representative breast models in order to efiectively compare the performance of each. This paper seeks to investigate the following beamforming algorithms: Monostatic and Multistatic Delay- And-Sum (DAS), Delay-Multiply-And-Sum (DMAS) and Improved Delay-And-Sum (IDAS). The performance of each beamformer is evaluated across a range of appropriate metrics.

Review: Audio quality assessment techniques-A review, and recent developments

Assessing the perceptual quality of wideband audio signals is an important consideration in many audio and multimedia networks and devices. Examples of such multimedia technologies are: streaming audio over the Internet, Digital Radio Mondiale (DRM), Digital Audio Broadcasting (DAB), VoIP (Voice over Internet Protocol), mobile phones, as well as compression algorithms for digital audio. The International Telecommunications Union (ITU) standard for audio quality (BS.1387) is commonly referred to as perceptual evaluation of audio quality (PEAQ). PEAQ is currently the only available standardised method for the purpose of audio quality assessment. This paper includes a brief technical summary of the standardised PEAQ algorithm. Furthermore, this paper outlines recent advancements in the general area of audio quality assessment since the publication of the ITU standard, and discusses possible techniques, including some recent findings, that could be used to extend the applicability of PEAQ and improve the accuracy of the algorithm in assessing the quality of multimedia devices and systems.

Equidistant Fish-Eye Calibration and Rectification by Vanishing Point Extraction

In this paper, we describe a method to photogrammetrically estimate the intrinsic and extrinsic parameters of fish-eye cameras using the properties of equidistance perspective, particularly vanishing point estimation, with the aim of providing a rectified image for scene viewing applications. The estimated intrinsic parameters are the optical center and the fish-eye lensing parameter, and the extrinsic parameters are the rotations about the world axes relative to the checkerboard calibration diagram.

Compressed Sensing for Bioelectric Signals: A Review

This paper provides a comprehensive review of compressed sensing or compressive sampling (CS) in bioelectric signal compression applications. The aim is to provide a detailed analysis of the current trends in CS, focusing on the advantages and disadvantages in compressing different biosignals and its suitability for deployment in embedded hardware. Performance metrics such as percent root-mean-squared difference (PRD), signal-to-noise ratio (SNR), and power consumption are used to objectively quantify the capabilities of CS. Furthermore, CS is compared to state-of-the-art compression algorithms in compressing electrocardiogram (ECG) and electroencephalography (EEG) as examples of typical biosignals. The main technical challenges associated with CS are discussed along with the predicted future trends.

SUPPORT VECTOR MACHINES FOR THE CLASSIFICATION OF EARLY-STAGE BREAST CANCER BASED ON RADAR TARGET SIGNATURES

Microwave Imaging (MI) has been widely investigated as a method to detect early stage breast cancer based on the dielectric contrast between normal and cancerous breast tissue at microwave frequencies. Furthermore, classiflcation methods have been developed to difierentiate between malignant and benign tumours. To successfully classify tumours using Ultra Wideband (UWB) radar, other features have to be examined other than simply the dielectric contrast between benign and malignant tumours, as contrast alone has been shown to be insuflcient. In this context, previous studies have investigated the use of the Radar Target Signature (RTS) of tumours to give valuable information about the size, shape and surface texture. In this study, a novel classiflcation method is examined, using Principal Component Analysis (PCA) to extract the most important tumour features from the RTS. Support Vector Machines (SVM) are then applied to the principal components as a method of classifying these tumours. Finally, several difierent classiflcation architectures are compared. In this study, the performance of classiflers is tested using a database of 352 tumour models, comprising four difierent sizes and shapes, using the cross validation method.