Research Interests:
Research Interests:
Research Interests:
Research Interests:
... 13 David Windridge, Vadim Mottl, Alexander Tatarchuk, and Andrey Eliseyev Kernel Combination Versus Classifier Combination ... 62 Sarunas Raudys, Omer Kaan Baykan, Ahmet Babalik, VitalijDenisov, and Antanas Andrius Bielskis Multiple... more
... 13 David Windridge, Vadim Mottl, Alexander Tatarchuk, and Andrey Eliseyev Kernel Combination Versus Classifier Combination ... 62 Sarunas Raudys, Omer Kaan Baykan, Ahmet Babalik, VitalijDenisov, and Antanas Andrius Bielskis Multiple Classifier Methods for Offline ...
Research Interests:
The Prague texture segmentation data-generator and benchmark (mosaic.utia.cas.cz) is a web-based service designed to mutually compare and rank (recently nearly 200) different static and dynamic texture and image segmenters, to find... more
The Prague texture segmentation data-generator and benchmark (mosaic.utia.cas.cz) is a web-based service designed to mutually compare and rank (recently nearly 200) different static and dynamic texture and image segmenters, to find optimal parametrization of a segmenter and support the development of new segmentation and classification methods. The benchmark verifies segmenter performance characteristics on potentially unlimited monospectral, multispectral, satellite, and bidirectional texture function (BTF) data using an extensive set of over forty prevalent criteria. It also enables us to test for noise robustness and scale, rotation, or illumination invariance. It can be used in other applications, such as feature selection, image compression, query by pictorial example, etc.The benchmark's functionalities are demonstrated in evaluating several examples of leading previously published unsupervised and supervised image segmentation algorithms. However, they are used to illustrate the benchmark functionality and not review the recent image segmentation state-of-the-art.
Research Interests:
A new unsupervised multispectral texture segmentation method with unknown number of classes is presented. Mul-tispectral texture mosaics are locally represented by four causal multispectral random field models recursively evalu-ated for... more
A new unsupervised multispectral texture segmentation method with unknown number of classes is presented. Mul-tispectral texture mosaics are locally represented by four causal multispectral random field models recursively evalu-ated for each pixel. The segmentation algorithm is ...
Research Interests:
Research Interests:
Visual texture fidelity evaluation is important but still unsolved problem. Evaluation of how well various texture models conform with human visual perception of their original measured pattern is required not only for assessing the... more
Visual texture fidelity evaluation is important but still unsolved problem. Evaluation of how well various texture models conform with human visual perception of their original measured pattern is required not only for assessing the visual dissimilarities between a model output and the original measured texture, but also for optimal settings of model parameters, for fair comparison of distinct models, or visual scene understanding. We propose a novel texture fidelity criterion based on the fully multi-spectral generative underlying Markovian texture model, which correlates well with human texture quality ranking verified on the texture fidelity benchmark.
Research Interests:
A novel unsupervised multi-spectral multiple-segmenter texture segmentation method with unknown number of classes is presented. The unsupervised segmenter is based on a combination of several unsupervised segmentation results, each in... more
A novel unsupervised multi-spectral multiple-segmenter texture segmentation method with unknown number of classes is presented. The unsupervised segmenter is based on a combination of several unsupervised segmentation results, each in different resolution, using the sum rule. Multi-spectral texture mosaics are locally represented by four causal multi-spectral random field models recursively evaluated for each pixel. The single-resolution segmentation part of the
Research Interests:
Research Interests:
Research Interests:
Research Interests: Geography, Mathematics, Computer Science, Artificial Intelligence, Computer Vision, and 13 moreEye tracking, Key words, Visual Performance, Visual, Degradation, Human Vision, Texture, Texture Features, Visual Psychophysics, Surface Texture, Representation Politics, Automatic Detection, and Statistical Features
Research Interests:
Research Interests:
Research Interests: Mathematics, Applied Mathematics, Computer Science, Pattern Recognition, Data Compression, and 11 moreVirtual Reality, Cluster Analysis, Computer Mathematics, Texture Synthesis, Spatial Clustering, Indexation, Spatial Dependence, Texture Mapping, Functional Dependency, Markov Random Field, and Springer Ebooks
This paper introduces multispectral rotationally in-variant textural features of the Markovian type applied for the effective coniferous tree needles categorization. Presented texture features are inferred from the descriptive... more
This paper introduces multispectral rotationally in-variant textural features of the Markovian type applied for the effective coniferous tree needles categorization. Presented texture features are inferred from the descriptive multispectral spiral wide-sense Markov model. Unlike the alternative texture recognition methods based on various gray-scale discriminative textural descriptions, we take advantage of the needles texture representation, which is fully descriptive multispectral and rotationally invariant.The presented method achieves high accuracy for needles recognition. Thus it can be used for reliable coniferous tree taxon classification. Our classifier is tested on the open source needles database Aff, which contains 716 high-resolution images from 11 diverse coniferous tree species.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Museums and other cultural heritage custodians are interested in digitizing their collections, not only for the sake of preserving cultural heritage, but also to make the information content accessible and affordable to researchers and... more
Museums and other cultural heritage custodians are interested in digitizing their collections, not only for the sake of preserving cultural heritage, but also to make the information content accessible and affordable to researchers and the general public. Once an objects digital model is created it can be digitally reconstructed to its original uneroded or unbroken shape or realistically visualized using different historical materials. Some artifacts are so fragile that they cannot leave the carefully controlled light, humidity, and temperature of their storage facilities, thus they are already inaccessible to the public, and the viable alternative is their exhibition in the form of an augmented reality scene. We present a sophisticated measurement and processing setup, which we have developed, to enable the construction of physically correct virtual models. This setup is illustrated on the reconstruction of one of the best known Celtic artifact from the European Iron Age period to its original uncorrupted form.
Research Interests:
Research Interests:
Research Interests:
Research Interests:
This paper introduces a novel multidimensional statistical model for realistic modeling, enlargement, editing, and compression of the recent state-of-the-art bidirectional texture function (BTF) textural representation. The presented... more
This paper introduces a novel multidimensional statistical model for realistic modeling, enlargement, editing, and compression of the recent state-of-the-art bidirectional texture function (BTF) textural representation. The presented multispectral compound Markov random field model (CMRF) efficiently fuses a non-parametric random field model with several parametric random fields models. The primary purpose of our modeling texture approach is to reproduce, compress, and enlarge a given measured natural or artificial texture image so that ideally both natural and synthetic texture will be visually indiscernible for any observation or illumination directions. However, the model can be easily applied for BFT material texture editing as well. The CMRF model consists of several parametric sub-models each having different characteristics along with an underlying switching structure model which controls transitions between these sub models. The proposed model uses the non-parametric random field for distributing local texture models in the form of analytically solvable wide-sense BTF Markov representation for single regions among the fields of a mosaic approximated by the random field structure model. The non-parametric control field of BTF -CMRF is reiteratively generated to guarantee identical region-size histograms for all material sub-classes present in the target example texture. The local texture regions (not necessarily continuous) are represented by analytical BTF models modeled by the adaptive 3D causal auto-regressive (3DCAR) random field model which can be analytically estimated as well as synthesized. The visual quality of the resulting complex synthetic textures generally surpasses the outputs of the previously published simpler non-compound BTF -MRF models. The model allows reaching huge compression ratio incomparable with any standard image compression method.
Research Interests:
Visual texture modeling based on multidimensional mathematical models is the prerequisite for both robust material recognition as well as for image restoration, compression or numerous physically correct virtual reality applications. A... more
Visual texture modeling based on multidimensional mathematical models is the prerequisite for both robust material recognition as well as for image restoration, compression or numerous physically correct virtual reality applications. A novel multispectral visual texture modeling method based on a descriptive, unusually complex, three-dimensional, spatial Gaussian mixture model is presented. Texture synthesis benefits from easy computation of arbitrary conditional distributions from the model. The model is inherently multispectral thus it does not suffer with the spectral quality compromises of the spectrally factorized alternative approaches. The model is especially well suited for multispectral textile textures and it can also describe the most advanced textural representation in the form of a bidirectional texture function (BTF).
Research Interests:
Research Interests:
The paper presents a detailed study of surface material recognition dependence on the illumination and viewing conditions which is a hard challenge in a realistic scene interpretation. The results document sharp classification accuracy... more
The paper presents a detailed study of surface material recognition dependence on the illumination and viewing conditions which is a hard challenge in a realistic scene interpretation. The results document sharp classification accuracy decrease when using usual texture recognition approach, i.e., small learning set size and the vertical viewing and illumination angle which is a very inadequate representation of the enormous material appearance variability. The visual appearance of materials is considered in the state-of-the-art Bidirectional Texture Function (BTF) representation and measured using the upper-end BTF gonioreflectometer. The materials in this study are sixty-five different wood species. The supervised material recognition uses the shallow convolutional neural network (CNN) for the error analysis of angular dependency. We propose a Gaussian mixture model-based method for robust material segmentation.
Research Interests:
Research Interests:
This paper describes a simple method for seamless enlargement of natural bidirectional texture functions (BTF) that realistically represent appearance of given material surfaces. The novel texture synthesis method, which we call the BTF... more
This paper describes a simple method for seamless enlargement of natural bidirectional texture functions (BTF) that realistically represent appearance of given material surfaces. The novel texture synthesis method, which we call the BTF roller, is based on the overlapping tiling and subsequent minimum error boundary cut. One or several optimal double toroidal BTF patches are seamlessly repeated during the synthesis step. While the method allows only moderate texture compression it is extremely fast due to complete separation of the analytical step of the algorithm from the texture synthesis part. The method is universal and easily implementable in a graphical hardware for purpose of realtime rendering of any type of static textures. 1.
Research Interests:
This paper describes a method for synthesizing natural textures that realistically matches given colour texture appearance. The novel texture synthesis method, which we call the roller, is based on the overlapping tiling and subsequent... more
This paper describes a method for synthesizing natural textures that realistically matches given colour texture appearance. The novel texture synthesis method, which we call the roller, is based on the overlapping tiling and subsequent minimum error boundary cut. An optimal double toroidal patch is seamlessly repeated during the synthesis step. While the method allows only moderate texture compression it is extremely fast due to separation of the analytical step of the algorithm from the texture synthesis part, universal, and easily implementable in a graphical hardware for purpose of real-time colour texture rendering.
Research Interests:
An efficient bark recognition method based on a novel wide-sense Markov spiral model textural representation is presented. Unlike the alternative bark recognition methods based on various gray-scale discriminative textural descriptions,... more
An efficient bark recognition method based on a novel wide-sense Markov spiral model textural representation is presented. Unlike the alternative bark recognition methods based on various gray-scale discriminative textural descriptions, we benefit from fully descriptive color, rotationally invariant bark texture representation. The proposed method significantly outperforms the state-of-the-art bark recognition approaches in terms of the classification accuracy.
Research Interests:
We present a novel multiimage restoration method based on the underlying spatial probabilistic image model for astronomical image restoration if degradation obeys a linear degradation model with the unknown possibly nonhomogeneous... more
We present a novel multiimage restoration method based on the underlying spatial probabilistic image model for astronomical image restoration if degradation obeys a linear degradation model with the unknown possibly nonhomogeneous point-spread function. The method assumes that for every ideal undegraded unobservable image several degraded observed images are available. Pixels in the vicinity of image steep discontinuities are left unrestored to minimise restoration blurring effect. The degraded image is assumed to follow a causal simultaneous multidimensional regressive model which leads to an adaptive fast recursive restoration filter. The unknown point-spread function is estimated using the local least-square estimate.
Visualization helps us to understand single-label and multi-label classification problems. In this paper, we show several standard techniques for simultaneous visualization of samples, features and multi-classes on the basis of linear... more
Visualization helps us to understand single-label and multi-label classification problems. In this paper, we show several standard techniques for simultaneous visualization of samples, features and multi-classes on the basis of linear regression and matrix factorization. The experiment with two real-life multi-label datasets showed that such techniques are effective to know how labels are correlated to each other and how features are related to labels in a given multi-label classification problem.
Research Interests: Computer Science and IEEE
The paper presents a detailed study of surface material recognition dependence on the illumination and viewing conditions which is a hard challenge in a realistic scene interpretation. The results document sharp classification accuracy... more
The paper presents a detailed study of surface material recognition dependence on the illumination and viewing conditions which is a hard challenge in a realistic scene interpretation. The results document sharp classification accuracy decrease when using usual texture recognition approach, i.e., small learning set size and the vertical viewing and illumination angle which is a very inadequate representation of the enormous material appearance variability. The visual appearance of materials is considered in the state-of-the-art Bidirectional Texture Function (BTF) representation and measured using the upper-end BTF gonioreflectometer. The materials in this study are sixty-five different wood species. The supervised material recognition uses the shallow convolutional neural network (CNN) for the error analysis of angular dependency. We propose a Gaussian mixture model-based method for robust material segmentation.
Research Interests:
Fast novel texture spectral similarity criterion, capable of assessing spectral modeling resemblance of color and Bidirectional Texture Functions (BTF) textures, is presented. The criterion reliably compares the multi-spectral pixel... more
Fast novel texture spectral similarity criterion, capable of assessing spectral modeling resemblance of color and Bidirectional Texture Functions (BTF) textures, is presented. The criterion reliably compares the multi-spectral pixel values of two textures, and thus it allows to assist an optimal modeling or acquisition setup development by comparing the original data with its synthetic simulations. The suggested criterion, together with existing alternatives, is extensively tested in a long series of thousands specially designed monotonically degrading experiments moreover, successfully compared on a wide variety of color and BTF textures.
Research Interests:
The spur-thighed tortoise (Testudo graeca) is listed among endangered species on the CITES list and the need to keep track of its specimens calls for a noninvasive, reliable and fast method that would recognize individual tortoises one... more
The spur-thighed tortoise (Testudo graeca) is listed among endangered species on the CITES list and the need to keep track of its specimens calls for a noninvasive, reliable and fast method that would recognize individual tortoises one from another. We present an automatic system for the recognition of tortoise specimen based on variable-quality digital photographs of their plastrons using an image classification approach and our proposed discriminative features. The plastron image database, on which the recognition system was tested, consists of 276 low-quality images with a variable scene set-up and of 982 moderate-quality images with a fixed scene set-up. The achieved overall success rates of automatically identifying a tortoise in the database were 43,0% for the low-quality images and 60,7% for the moderate-quality images. The results show that the automatic tortoise recognition based on the plastron images is feasible and suggests further improvements for a real application use.
In this research report we present a new model for texture representation which is particularly well suited for image analysis and segmentation. Any image is first discretized and then a hierarchical finite-state region-based model is... more
In this research report we present a new model for texture representation which is particularly well suited for image analysis and segmentation. Any image is first discretized and then a hierarchical finite-state region-based model is automatically coupled with the data by means of a sequential optimization scheme, namely the Texture Fragmentation and Reconstruction (TFR) algorithm. The TFR algorithm allows to model both intra- and inter-texture interactions, and eventually addresses the segmentation task in a completely unsupervised manner. Moreover, it provides a hierarchical output, as the user may decide the scale at which the segmentation has to be given. Tests were carried out on both natural texture mosaics provided by the Prague Texture Segmentation Datagenerator Benchmark and remote-sensing data of forest areas provided by the French National Forest Inventory (IFN).
Research Interests: Geography, Computer Science, Artificial Intelligence, Image Analysis, Segmentation, and 10 moreClassification, Texture Synthesis, Texture Segmentation, Gray Level Co-occurrence Matrix, THERMAL INFRARED REMOTE SENSING DATA, Markov chain, INRIA, National forest inventory, Structural model, and remote sensing data
Research Interests:
Research Interests:
Research Interests:
To enhance realism in a graphics system it is necessary to cover generated surfaces with realistic textures. A texture in this paper is assumed to be a realization of a random field. Problems of parameter estimation and random field... more
To enhance realism in a graphics system it is necessary to cover generated surfaces with realistic textures. A texture in this paper is assumed to be a realization of a random field. Problems of parameter estimation and random field synthesis of a given random field model are studied. The most flexible models for realistic texture synthesis seem to be simultaneous autoregressive and Gaussian Markov random field models.