Boban P. Bondzulic

The database of radar echoes from various targets with spectral analysis

In this paper we describe a database, noted as RadEch Database, containing radar echoes from various targets. The data has been collected in controlled test environments at the premises of Military Academy — Republic of Serbia. Our goal is to provide a balanced and comprehensive database to enable reproducible research results in the field of classification of ground moving targets (pattern recognition). A time-frequency analysis of radar echoes has been performed, in order to identify the main features of the various targets. The RadEch Database is freely available for download and we hope that our database provides researchers with a valuable tool to benchmark and improve the performance of classification algorithms.

Edge-based objective evaluation of image quality

An objective metric for grayscale image quality capable of both global and localized quality assessment and based on a direct comparison of visual information in the test and reference images is proposed. The measure associates visual information with edge, or gradient, information that is initially extracted at all locations of the test and reference images. A perceptual-information preservation model is then used to quantify the success of information transfer/loss. By considering the perceptual importance of different image regions, local image quality success estimates are integrated into a single, numerical quality score between 0 (total information loss) and 1 (ideal transfer). The proposed metric is validated using extensive subjective test results. Results indicate that even in its basic form the proposed metric is perceptually meaningful and yields performance comparable with state-of-the art objective image quality metrics.

Additive models and separable pooling, a new look at structural similarity

Objective quality metrics predict perceived quality of image signals computationally and can: (i) benchmark and monitor compression and processing algorithms and (ii) optimise their performance for a given application (content, bandwidth, packet loss?). Structural similarity, represented by the well known SSIM index is a framework for objective assessment of image quality well known for its relative simplicity and robustness. Despite its practical appeal, SSIMs performance level, measured as agreement with subjective quality scores, lags more complex state-of-the-art metrics. We present a new look into structural similarity that uses an additive model and a spatial pooling approach that decouples individual structural comparisons and utilises the quality driven aggregation paradigm. We apply this new approach to both baseline intensity SSIM and gradient SSIM (GSSIM) frameworks and show, through extensive evaluation on four publicly available subjective datasets that it provides considerably more ordered (linear) relationship between objective and subjective quality for a variety of input conditions. We demonstrate that newly formulated structural similarity metrics using this approach are capable of equal or even better performance than more complex state-of-the-art objective metrics in the process lending support to the theory that humans base their opinion on the worst sections of the observed signal. We present a new look into structural similarity (SSIM) index.Newly formulated approach uses an additive model and a spatial pooling approach.It provides considerably linear relationship between objective and subjective quality.We show that using the proposed approach can improve SSIM performance.

Cepstrum-based analysis of radar Doppler signals

The main tasks of ground surveillance radars for security and perimeter protection are detection and classification of moving ground targets. In typical radar systems, target detection is fully automated, but the target classification requires human involvement. In this paper, we consider received radar echoes data of ground moving targets, and corresponding signals using cepstrum coefficients. The objective of the paper is to identify and validate features characterizing the different classes of targets, and subsequently extract features for classification. We will show examples on Radar Echoes Database. This database, named RadEch Database, contains radar echoes that are collected in controlled test enviroments at the premises of Military Academy — Republic of Serbia. Database purpose is to enable reproducible research results in the field of classification of ground moving targets (pattern recognition).

Expert system based on hidden Markov models for recognition of radar targets

Design of an expert system based on Hidden Markov Models for recognition of radar targets in a zone of ground surveillance radar is presented in the paper. Parameters of the real radar echo signal represented in a form of autoregressive models are used as an input of the designed expert system. The real radar echoes have been collected for the purpose of this research. Obtained results show that designed system has some certain advantages, but there are also some limitations in recognition of the analyzed sequences.

Objective image quality characterisation

This paper proposes a method for comprehensive, objective image quality characterisation using an image quality evaluation framework based on gradient information representation. The method provides an in-depth analysis of image quality by quantifying: common information, information loss and information artifacts. The results demonstrate and quantify a number of well known issues and provide a useful insight into a no-reference image quality assessment.