Vol. 20 (2011):
  Abstracts of Papers

19 (2010)	main	21 (2012)

No. 1,
No. 2,
No. 3,
No. 4.

Machine GRAPHICS & VISION, Vol. 20 (2011), No. 1:

Darles E., Crespin B., Ghazanfarpour D.:
A Particle-Based Method for Large-Scale Breaking Wave Simulation.
MGV vol. 20, no. 1, 2011, pp. 3-25.

In this paper we address the problem of particle-based simulation of breaking waves. We present a new set of equations based on oceanographic research which allow us to deal with several types of breaking waves and multiple wave trains with full control over the governing parameters. Sprays are generated by explicitly computing sub-particle systems depending on the local motion caused by plunging. In order to reduce computations in non-significant areas, we also describe a simple and efficient multiresolution scheme based on the properties of our breaking wave model.
Key words: Fluid simulation, breaking waves, smooth particle hydrodynamics, multiresolution.

Pamuła W.:
Wavelet-Based Data Reduction for Detection of Moving Objects.
MGV vol. 20, no. 1, 2011, pp. 27-40.

The detection of moving objects in video streams is generally performed by analysis of the differences between the modelled background and the current stream content, by matching object models, extracting and clustering the features of objects or else by using various filtering methods. Filtering is performed on the transformed contents of the video stream. Due to implementational constraints, mainly limited processing resources, solutions based on these principles of detection are sensitive to ambient light variations, objects shadows and camera movement. This paper presents a method for the detection of moving objects that uses a data reduction technique based on wavelets. Instead of the analysis of raw video data, wavelet coe#cients of an appropriate scale are explored. In order to satisfy low processing requirements, an integer version of discrete wavelet transform is chosen for processing. To facilitate the detection, each frame is converted into a vector of pixel values. Consecutive video vectors are transformed using one-dimensional Discrete Wave Transform (DWT). The computed DWT coe#cients make up a surface, which maps changes in their values over time. The surface is analysed to find clusters of values corresponding to moving objects. The checked patches represent moving objects. The width of a patch indicates the object size. Background details and illumination changes are represented by gradually changing patterns. Various examples demonstrate the potential of the method for practical applications.
Key words: Video stream, DWT, moving objects, analysis of temporal changes.

Rodríguez T.:
A Novel Approach to Rail Crossing Protection Using Computer Vision and Radio Communications.
MGV vol. 20, no. 1, 2011, pp. 41-71.

In the paper we present an innovative computer vision based rail crossing protection system. A camera installed on top of a mast overlooking the crossing continuously monitors the scene, searching for objects that have stopped on the rail tracks. The system is designed to transmit images of the incident to the approaching trains as soon as any conflictive object has been detected. A simple user interface on board of the train displays the image sequence with a graphical aid clearly identifying the o#ending object in the image. In that way, train drivers are alerted of the presence of possible obstacles well before the train has approached the crossing. The system we describe operates autonomously for long periods of time without human intervention and adapts automatically to the changing environmental conditions. Several innovations, designed to deal with the above circumstances, are proposed in the paper, including: an adaptive segmentation algorithm, an innovative method for the detection of stopped objects and di#erentiated approaches for day and night processing.
Key words: Rail crossing, real-time, computer vision, background substraction, non-motion detection, segmentation, object detection and tracking.

Goszczyńska H., Doros M., Kowalczyk L., Zalewska E., Kolebska K.:
Detection of similar sequences in EEG maps series using correlation coefficients matrix.
MGV vol. 20, no. 1, 2011, pp. 73-92.

The aim of this study has been to develop a method to indicate the similar sequences of electroencephalographic (EEG) maps in a series. A method for the analysis of sequence similarity using the matrix of correlation coefficients for each pair of the EEG maps in the series has been proposed. The results for two series of EEG maps for seizure activity episodes and for activity before, during and after the seizure episode are presented. Analysis of images of the correlation coe#cients matrices has allowed us to determine the characteristic features of the areas in these matrices corresponding to the assumed similarity relations, and to indicate the sequences fulfilling these relationships.
Key words: Image sequence similarity measures, correlation matrix, EEG maps sequences.

Abdallah M.M.:
ANN Face Detection with Skin Color Distribution Rules.
MGV vol. 20, no. 1, 2011, pp. 93-108.

This paper develops a face detection method in color images using a multi-layer Neural Network classification. The proposed method is based on two image processing steps which first de- tect skin regions in the color image and then extract face information from those regions. Instead of performing huge search in every part of the test images, a pre-processing method for candidate face regions guides the image search using neural networks. The new algorithms perform fast and accurate face detection. Experiments have been carried out and satisfactory results have been obtained which indicate the robustness of the first process to detect faces under different environmental conditions.
Key words: Skin detection, face, Artificial Neural Networks.

New Books Notes
MGV vol. 20, no. 1, 2011, pp. 109-110.

Ryszard Tadeusiewicz, Jacek Śmietański: Medical Image Aquisition Processing, Analysis and Diagnostic Interpretation (in Polish).
Published by Scientific Society Publishers Cracow 2011.

Machine GRAPHICS & VISION, Vol. 20 (2011), No. 2:

Zawadzki T., Nikiel S., Warszawski K.:
Procedural Modelling of Three-Dimensional Geometry
MGV vol. 20, no. 2, 2011, pp. 113-138.

Modelling three-dimensional shapes plays an increasingly significant role in modern computer graphics. Geometry synthesis is used in many fields, including digital cinema, electronic entertainment and computer simulations. Unfortunately, the modelling process is still done manually, offering a unique output at the cost of tedious work. There is a constant need to replace designers' work with intelligent automated algorithms. The methods based on the automation of modelling processes offer a variety of three-dimensional structures within limited time and restricted money budget. This paper addresses the problem of automated modelling of virtual structures such as caves, buildings and clouds, and presents an alternative solution in the form of a hybrid system. The innovative approach combines two independent methods well known in three-dimensional computer graphics: shape grammar and shape morphing. In the modelling process, it is possible to obtain the characteristics of 3D structures with non-spherical mesh topology. The objects and their transformations are described by functions, while rule grammars define the geometry modelling process. The shapes thus obtained can be freely deformed in the subsequent rules. The resulting structure can be rendered up to very high levels of visual realism. However, in the paper we present the description of the algorithm illustrated by results on a 3D mesh without focusing on photorealistic rendering aspects. We also propose some measures that can be used to verify the model geometry.
Key words: Shape grammar, morphing, procedural modelling, hybrid, polygon.

Szajerman D., Pełczyński P.:
Efficient implementation of stereovision algorithms for graphics processing unit in DirectShow technology
MGV vol. 20, no. 2, 2011, pp. 139-

Stereovision is a passive technique for estimation of depth in 3D scenes. Unfortunately, depth estimation in this imaging technique is computationally demanding. We show that stereovision matching algorithms can be efficiently mapped onto the present-day graphics processing units (GPUs). A number of modifications to the original image disparity estimation algorithm have been proposed that make running its computation on GPU platforms particularly efficient. A complete depth estimation system was implemented in GPU, covering correction of camera distortions, image rectification and disparity estimation. To obtain modularity of developed software, the DirectShow multimedia technology was used. Examples, computed depth maps are shown, and time performances of the proposed algorithms are outlined. The developed system has proved the usefulness of both GPU implementation and the DirectShow technology in scene depth estimation.
Key words: Stereovision, depth estimation, Graphics Processing Unit, DirectShow.

Bezin R., Peyrat A., Crespin B., Terraz O., Skapin X., Meseure P.:
Interactive Hydraulic Erosion Using CUDA
MGV vol. 20, no. 2, 2011, pp. 157-172

This paper presents a method to simulate in real-time hydraulic erosion and sedimentation on a 3D soil represented by a triangular mesh. Applications include user-assisted terrain generation for computer-generated films and reverse engineering for geology. Our method achieves interactive performances by dynamically displacing vertices using CUDA and following physically-inspired principles to simulate realistic water streams and their interaction with soil and sediments. The mesh is generated in a preprocessing step to avoid degenerate cases in highly deformed areas. We present various results where landforms are progressively eroded to create visually plausible river beds.
Key words: Computer Graphics, Automatic Terrain Generation, Hydraulic Erosion, Fluid Simulation.

R. Todmal Satish, K. Thammi Reddy:
A Technique to Find Optimal Location for Wavelet-Based Image Watermarking Using Genetic Algorithm
MGV vol. 20, no. 2, 2011, pp. 173-196

In this paper we developed an efficient optimal robust watermarking technique using genetic algorithm (GA) for images of Indian historical monuments and their corresponding names. The watermarks are embedded into the HL and LH frequency coefficients in the Haar wavelet transform domain. Since the embedding technique is blind, it does not require the original image in the watermark extraction. We also develop an optimization technique using the GA to search for the optimal locations in order to improve both quality of watermarked image and robustness of the watermark. We analyze the performance of the proposed watermarking technique in terms of peak signal-to-noise ratio (PSNR) and normalized correlation (NC). The experimental and the comparative results show that the proposed technique can achieve a good robustness against most of the attacks which are included in this study. For typical image quality, the proposed technique outperforms the existing one with a PSNR of 36 dB and the NC value of 0.96.
Key words: Digital image watermarking, Haar wavelet transform, GA, Indian historical monument images, robustness, PSNR, NC.

Kumar S.S., Moni R.S., Rajeesh J.:
Liver Tumour Classification Using Co-occurrence Matrices on the Contourlet Domain
MGV vol. 20, no. 2, 2011, pp. 197-214

Liver disease is one of the most common diseases around the world, seriously affecting the health of humans. Computed tomography image based Computer Aided Diagnosis (CAD) could be crucially important in supporting liver cancer diagnosis. An effective approach to realize a CAD system for this purpose is described in this work. The CAD system employs automatic tumour segmentation, texture feature extraction and characterization into malignant and benign tumours. A Region of Interest (ROI) cropped from the automatically segmented tumour by confidence connected region growing and alternative fuzzy c means clustering is decomposed using multiresolution and multidirectional contourlet transform to obtain contourlet coefficients. Co-occurrence matrices of the contourlet coefficients are determined, and six parameters of texture characteristics, which include Angular Second Moment, Contrast, Correlation, Inverse Difference Moment, Entropy and Variance, are extracted from them. The extracted feature sets are classified into benign and malignant by a Generalized Regression Neural Network (GRNN) classifier. The performance of this scheme is evaluated by various performance measures and by the use a of the Receiver Operating Characteristic (ROC) curve. The results are compared with those obtained by a similar system using Wavelet Coefficients co-occurrence Matrix (WCCM) and Gray Level co-occurrence Matrix (GLCM) texture features. The results indicate that the proposed scheme based on the CCCM texture is effective for classifying malignant and benign liver tumours in abdominal CT imaging.
Key words: Liver Tumour, Texture Analysis, Co-occurrence Matrix, GLCM, Wavelet Transform, Contourlet Transform.

Kmieć M., Głowacz A.:
An Approach to Robust Visual Knife Detection
MGV vol. 20, no. 2, 2011, pp. 215-227

Computerised monitoring of CCTV images is attracting a lot of attention both from potential end-users seeking to increase the effectiveness of their video surveillance systems and as a popular research topic as new methods and algorithms are being developed. In this paper an approach to detecting knives in images is presented. It is based on the use of Histograms of Oriented Gradients (HOG), feature descriptors invariant to geometric and photometric transformations except for rotation. We introduce a dataset containing images of knives in different backgrounds and in varying lighting conditions and evaluate the performance of an HOG-based SVM classifier. We study the question of creating a detector based on knife blade colour and discuss the use of GPU parallel computing as a method of speeding up the detection process.
Key words: Knife Detection, Dangerous Tool Detection, Histograms of Oriented Gradients, Computerised Monitoring, Parallel Computing, GPU.

MGV vol. 20, no. 2, 2011, pp. 216
The 22^nd IFIP World Computer Congress 2012 Announcement

Machine GRAPHICS & VISION, Vol. 20 (2011), No. 3:

Szczesna A., Słupik J., Janiak M.:
Motion Data Denoising Based on the Quaternion Lifting Scheme Multiresolution Transform.
MGV vol. 20, no. 3, 2011, pp. 237-249.

We discuss human body motion denoising with use of the transform based on second generation wavelet. To build such a transform, we use the quaternion lifting scheme. The main focus is placed on representing body parts orientation changes over time with quaternions as a technique both compact and more efficient than the representation of the Euler angles of rotation. Our denoising method is based on a soft threshold algorithm but is directly to the quaternion motion data in the resulting multiresolution representation.
Key words: Quaternions, multiresolution analysis, wavelet transform, lifting scheme, motion data denoising.

Jabłoński B., Kulbacki M.:
Multiscale Processing Performance for Motion Capture.
MGV vol. 20, no. 3, 2011, pp. 251266.

Motion capture systems help record human motion as a sequence of joint angle vectors and analyse it in multiple degrees of freedom with high accuracy. Motion, as many other signals, might contain information which is stored on many different scales. Hence the use of a multiscale model might help correctly distinguish or analyse motion properties. In this paper we analyse the capabilities of a multiscale motion model to help distinguish meaningful motionf eatures, whilst the unnecessary components (like noise) get removed. We performed experiments based on real motion capture data to analyse the discriminative properties of the multiscale approach. The main goal of experiments was to check the clustering performance of a multiscale model. The detailed results are presented and discussed, showing the capabilities and advantages of multiscale model application.
Key words: Human motion, multiscale, motion processing, performance analysis.

Krzeszowski T., Kwolek B., Wojciechowski K., Josiński H.:
Markerless Articulated Human Body Tracking for Gait Analysis and Recognition.
MGV vol. 20, no. 3, 2011, pp. 267281.

We present a particle swarm optimization (PSO) based system for markerless full body motion tracking. The fitness function is smoothed in an annealing scheme and then quantized. In this manner we extract a pool of candidate best particles. The swarm of particles selects a global best from such a pool of the particles to force the PSO the jump out of stagnation. Experiments on 4-camera datasets demonstrate the accuracy of our method on image sequences with walking persons. The system was evaluated using ground-truth data from a marker-based motion capture system by Vicon. We compared the joint motions and the distances between ankles, which were extracted using both systems. Thanks to the high precision of the markerless motion estimation, the curves illustrating the distances between ankles overlap considerably in almost all frames of the image sequences.
Key words: Markerless motion tracking, particle swarm optimization, gait analysis and recognition.

Stępień J., Lebek K.:
Exploiting QuTEM for Improving Joint Angles Estimation in the Context of Clinical Motion Analysis.
MGV vol. 20, no. 3, 2011, pp. 283-298.

This paper describes a statistical procedure designed to work with quaternions (QuTEM) as a useful tool for interpreting recorded joint motion in the particular context of clinical gait analysis. We believe that QuTEM addresses some of the problems of analyzing and interpreting joint motion data acquired using numerous available devices. The theoretical introduction is followed by a measured noise robustness test and two usage scenarios describing actual practical applications of this procedure.
Key words: Principal Component Analysis, Data Dimensionality Reduction, Inverse Dynamics, Human Gait, Quaternions, QuTEM, Joint Models.

Stawarz M., Kwiek S., Polański A., Janik Ł., Boczarska-Jedynak M., Przybyszewski A., Wojciechowski K.:
Algorithms for Computing Indexes of Neurological Gait Abnormalities in Patients after DBSS urgery for Parkinson Disease Based on Motion Capture Data.
MGV vol. 20, no. 3, 2011, pp. 299-317.

The motion capture (MOCAP) technology is becoming very useful in medical applications, such as orthopedics, neurology and physical therapy. In the literature the reare already studies of using MOCAP measurements for diagnosis of gait abnormalities. In this paper, we present an algorithm for computation of important indexes defning abnormalities of the gait for a unifed MOCAP data standard, namely for the ASF/AMCMOCAP data recording and storage system. We also present preliminary experimental results on examination of Parkinson's disease (PD) for a bilateral subthalamic nucleus stimulation patient in the MOCAP laboratory.
Key words: Parkinson Disease, Motion Capture, Gait Analysis.

Dlugosz R., Pauk J., Farine P.-A.:
New Trends in Motion Capture Systems for Human Gait Analysis.
MGV vol. 20, no. 3, 2011, pp. 319-331.

The paper presents new trends in motion capture systems for human gait analysis, such as: resonant magnetic techniques, inertial techniques, and wireless positioning techniques based on Ultra-Wideband (UWB) communication. The new techniques, due to the irrelatively low expected prices, can replace expensive optical systems that are commonly used today. As the new systems are mostly based on the Wireless Sensor Network (WSN) technology, any improvement in their ranging precision, which is a key aspect in this case requires development of new ultra low power microelectronics circuits. Selected solutions in this area are presented in the paper.
Key words: Motion capture, ASIC, ADC, MEMS gyroscopes, Decimation Flter, Wireless Sensor Networks.

Janik Ł., Polański A., Wojciechowski K.:
Models and Algorithms for Human Skeleton Estimation from 3D Marker Trajectories.
MGV vol. 20, no. 3, 2011, pp. 333-354.

The optical motion capture (MOCAP) technique is very useful in numerous applications in technology, medicine and science. The central problem in the MOCAP technology is skeleton estimation. The quality of skeleton estimation algorithms has the most signifcant infuence on the whole process of MOCAP data processing and analysis.
In this paper we describe an original algorithm for skeleton estimation which refines several methodologies available in the literature. We have performed extensive testing and accuracy comparisons of our algorithm compared to some algorithms presented in the literature. We can report that our algorithm enables estimation of a high quality skeleton, with better accuracy than previous methods.
Key words: Skeleton Estimation, Motion Capture.

Pełczyński P.:
Disparity Modeling by Second-degree Surfaces for Unsupervised Segmentation of Stereoscopic Image Sequences.
MGV vol. 20, no. 3, 2011, pp. 355-377.

The paper describes a three-dimensional scene analysis technique, the goal of which is the segmentation of a sequence of images obtained from a stereoscopic camera system. The proposed scheme allows for continuous detection of new objects and tracking of those that have already been detected. Segmentation is the result of an algorithm that classifes disparity image points into one of a number of model surfaces. Objects are modelled with a set of second-degree surfaces described in the depth image space. The developed algorithm is to beimplemented in an obstacle avoidance aid for visually impaired persons.
Key words: Stereovision, Disparity, Second-degree Surface, Object Modelling, Image Segmentation.

New Books Notes
MGV vol. 20, no. 3, 2011, pp. 379-380.

Wojciech S. Mokrzycki: Introduction to Visual Information Processing. II. Image Discretization, Pixel Operations, Morphological Operations and Image Processing (in Polish).
Published by AOW EXIT, Warsaw 2012.

Machine GRAPHICS & VISION, Vol. 20 (2011), No. 4:

Mokrzycki WS., Tatol M.:
Color difference ΔE - A survey.
MGV vol. 20, no. 4, 2011, pp. 383-411.

Color perception is crucial for human existence. This is why, color spaces have been developed to describe mathematically the color that a person can feel with unaided eye. There was a new need to distinguish colors, define them as similar, identical or completely different. However, a color-matching technique requires a color palette with perceptually linear characteristics. In this articl the most popular colors spaces are presented, as both linear and nonlinear due to perceptual abilities, and are briefly discussed and compared to the sample values.
Key words: Color difference, ΔE, perceptual color spaces.

Denisiuk A.:
On numerical reconstruction of a function from incomplete data of arc means in seismic tomography.
MGV vol. 20, no. 4, 2011, pp. 413437.

Consider the problem of reconstruction of a small perturbation of the acoustic wave speed field from traveltime data with linear background slowness. Mathematically, the problem is equivalent to reconstruction of a function from the data of integrals along the circle arcs. The data is limited, in the sense that the base points belong to a compact set. We propose and numerically test a new approach, based on reduction of the problem to the inverse problem for the Radon transform. The data completion procedure is considered as well.
Key words: Radon transform, Seismic tomography, Inverse kinematic problem, Spherical mean transform, Interpolation of band-limited function.

Tarko J., Grabska E.:
Aesthetic measure for three-dimensional objects.
MGV vol. 20, no. 4, 2011, pp. 439454.

This paper discusses a possible extension of Birkhoff's aesthetic measure in order to apply it to three-dimensional objects. Modified definitions of complexity and order, analogous to their two-dimensional equivalents, are proposed and used for calculation of aesthetic measure for both abstract and real-world three-dimensional objects.
Key words: Computational aesthetics, aesthetic measure, complexity, order.

Przytulska M., Kulikowski J. L., Wierzbicka D.:
Discrimination of Poorly Distinguishable Random Textures by Statistical Analysis of Morphological Spectra .
MGV vol. 20, no. 4, 2011, pp. 455-477.

The paper describes a method for discrimination of poorly distinguishable textures based on application of morphological spectra. The textures are analyzed as random fields of specific probability distributions. The samples of textures are thus considered as their instances and so are also their morphological spectra. Some basic properties of morphological spectra, as well as the definition of similarity measure are shortly reminded. The problem of textures discrimination is formulated as similarity assessment of spectral components histograms. For this purpose, various statistics like: mean value, standard deviation, skewness and kurtosis, as well as some secondary statistics based on the former, are used. A discriminating index is introduced for evaluation of their discriminating properties. The method of evaluating the discriminating power of statistics based on 1^st and 2^nd level morphological spectra is illustrated by analysis of the spectra of USG liver images in the groups of healthy persons and patients affected by liver fibrosis. A short description of a IASS program used to the calculations is given. The problem of textures discrimination invariant to rotations and parallel translations of images is described. It is shown that the proposed method discriminates statistically the "ill" and "healthy" textures despite the fact that the differences between them are visually not distinguishable.
Key words: Computer-aided image analysis, textures discrimination, morphological spectra, statistical data analysis, USG liver imaging.

Juránek R., Machalik S., Zemcik P.:
Research of Image Features for Classification of Wear Debris.
MGV vol. 20, no. 4, 2011, pp. 479-493.

The wear debris of engineering equipment (such as combustion engines, gearboxes, etc.) consists of metal particles which can be obtained from lubricants used in the equipment. The analysis of wear particles is very important for early detection and prevention of failures. The analysis is often done using classication of individual wear particles obtained by analytical ferrography. In this paper, we present a study of feature extraction methods for a classication of wear particles based on visual similarity. The main contribution of the paper is the comparison of nine selected feature types in the context of three state-of-the-art learning models. Another contribution is the large public database of particle images which can be used for further experiments. The paper describes the dataset, presents the methods of classication, demonstrates the experimental results, and draws conclusions.
Key words: Wear Debris, Classification, Supervised Machine Learning, SVM, Linear Regression, Features, PCA, HOG, LBP.

Kornuta T., Bem T., Winiarski T.:
New Trends in Motion Capture Systems for Human Gait Analysis.
MGV vol. 20, no. 4, 2011, pp. 495-520.

The paper presents FraDIA, a framework facilitating the creation of vision systems, that can operate as a standalone application as well as play a role of a vision subsystem for robotic controllers. The article describes motivations leading to the tool creation, its structure and a method of integration with a MRROC++ system, enabling the development of a robot controllers with visual feedback. The usefulness of the framework is demonstrated on the example of a robot playing checkers. In the application, FraDIA was used to implement two different vision subsystems, and the control system exhibited two behaviors utilizing visual information in two totally different ways: passive, responsible for monitoring the state of the game, and active, in which vision was utilized during the manipulator motion for localization of a pawn to be grasped. Regarding the complexity of the system, a specification method based on agents and transition function was used. The method, consisting of mathematical formulas supplemented by data flow diagrams, enables the reader to understand both the system structure and its behavior.
Key words: Vision subsystem, robotic controller, behavior-based control, checkers game.

19 (2010)	main	21 (2012)