No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Abstract
The investigations on texture classification goes back to many years ago, but non of them has been succesfull to find out road edges due to complexness in texture nature and its inefficient processing algoritms This paper presents a new approach to determine obstacles on a road for autonomous car by dividing the road into regions of different texture patterns. Obstacles usually introduce some different texture patterns to an abstract road images. The patterns are recognized by the proposed method called "texture dissimilarity measure (DBÖ)", thus providing a safe driving.
ResearchGate has not been able to resolve any citations for this publication.
This paper presents a computational model that segments images based on the textural properties of object surfaces. The proposed Coupled-Membrane model applies the weak membrane approach to an im- age Wl(O' y), derived from the power responses of a family of self- similar quadrature Gabor wavelets. While segmentation breaks are allowed in x and y only, coupling is introduced to in all 4 dimensions. The result- ing spatial and spectral diffusion prevents minor variations in local tex- tures from producing segmentation boundaries. Experiments showed that the model is adequate in segmenting a class of synthetic and natural texture images.
Perception of different textures is caused by differences in distribution of properties of texture elements. However, in practice it is difficult to extract useful texture elements, especially from natural images in which texture elements exist at various scales. To extract texture elements of all sizes a multiscale approach is unavoidable. This paper describes a multiscale method, based on measurements in a Laplacian-of- Gaussian scale-space, to extract texture elements. Histograms are used to describe the distribution of properties of extracted texture elements in a region. The edge significance at a pixel reflects the difference in the histograms of the regions surrounding the pixel. High edge significance pixels constitute the texture boundaries. Performance of the approach is shown for various natural textured images.
This paper presents a neural-network model that is capable of recognising roads in colour images of urban and rural road scenes. The network is trained to classify the visible objects in an image, given a set of hand-labelled examples. The two problems of segmentation and recognition are separated through the use of “ideal” segmentations, allowing the performance of the recognition method to be assessed independently of the effects of using an imperfect real segmentation process. The generalisation performance of the method is measured using unseen test data. Results for both on-road and offroad viewpoints are presented. The system is found to recognise correctly approx. 70% of the regions, and approx. 90% of the area of the images
We address the problem of scale selection in texture analysis. Two different scale parameters, feature scale and statistical scale, are defined. Statistical scale is the size of the regions used to compute averages. We define the class of homogeneous random functions as a model of texture. A dishomogeneity function is defined and we prove that it has useful asymptotic properties in the limit of infinite statistical scale. We describe an algorithm for image partitioning which has performed well on piecewise homogeneous synthetic images. This algorithm is embedded in a redundant pyramid and does not require any ad-hoc information. It selects the optimal statistical scale at each location in the image.
This paper presents an edge finder for textured images. Using rough constraints on the size of image regions, it estimates the local amount of variation in image values. These estimates are constructed so that they do not rise at boundaries. This enables subsequent smoothing and edge detection to find coarse-scale boundaries to the full available resolution, while ignoring changes within uniformly textured regions. This method extends easily to vector valued images, e.g. 3-color images or texture features. Significant groups of outlier values are also identified, enabling the edge finder to detect cracks separating regions as well as certain changes in texture phase.
In this paper we define six parameters addressed to parametrize the texture characteristics of an image towards its segmentation. With the aim to operate at high speed, these parameters have been defined looking for an acceptable compromise between discrimination capacity and easyness to implement a specific architecture for them.
We present a new algorithm for segmentation of textured images
using a multiresolution Bayesian approach. The new algorithm uses a
multiresolution Gaussian autoregressive (MGAR) model for the pyramid
representation of the observed image, and assumes a multiscale Markov
random field model for the class label pyramid. The models used in this
paper incorporate correlations between different levels of both the
observed image pyramid and the class label pyramid. The criterion used
for segmentation is the minimization of the expected value of the number
of misclassified nodes in the multiresolution lattice. The estimate
which satisfies this criterion is referred to as the
“multiresolution maximization of the posterior marginals”
(MMPM) estimate, and is a natural extension of the single-resolution
“maximization of the posterior marginals” (MPM) estimate.
Previous multiresolution segmentation techniques have been based on the
maximum a posterior (MAP) estimation criterion, which has been shown to
be less appropriate for segmentation than the MPM criterion. It is
assumed that the number of distinct textures in the observed image is
known. The parameters of the MGAR model-the means, prediction
coefficients, and prediction error variances of the different
textures-are unknown. A modified version of the expectation-maximization
(EM) algorithm is used to estimate these parameters. The parameters of
the Gibbs distribution for the label pyramid are assumed to be known.
Experimental results demonstrating the performance of the algorithm are
presented
Statistics of Neighbouring Grey Level Differences (NGLD-1) for
- B G Çetiner
B. G. Çetiner, " Statistics of Neighbouring Grey Level Differences (NGLD-1) for