Unsupervised colour image segmentation by low-level perceptual grouping

Segmentation is a low-level processing aimed at the partition of an image in disjoint regions, each one homogeneous with respect to some properties like intensity, texture, shape, etc. Such a task is needed in many high-level processing and applications in such diverse fields as remote-sensing, medical imaging, video coding, and industrial automation, just to name a few. In this work of thesis, we first deal with the problem of color based image segmentation, that is based on some form of homogeneity in the color/spectral properties of the image, and then move to the more complex task of texture based segmentation, where the aim is to recognize complex structures in the image which are typically non homogeneous in terms of spectral properties. Concerning color based segmentation, we focus on a specific class of Markov Random Field (MRF) probability models, namely the Tree-Structured Markov Random Fields (TS-MRF), recently proposed to provide a hierarchical multiscale description of images. Based on such a model, the unsupervised image segmentation is carried out by means of a sequence of nested class splits, where each class is modeled as a local binary MRF. As a first contribution, we first extend the model to generic tree structures removing the unnecessary binary constraint, and then devise a new TS-MRF unsupervised segmentation technique which improves upon the original algorithm by selecting a better tree structure and eliminating spurious classes. Such results are obtained by using the Mean-Shift procedure to estimate the number of pdf modes at each node, and to obtain a more reliable initial clustering for subsequent MRF optimization. To this end, we devise a new reliable and fast clustering algorithm based on the Mean-Shift technique, which makes use of a variable-bandwidth strategy based on the k-Nearest Neighbors (k-NN) technique, and is implemented with a speed-up strategy that cuts significantly the computational complexity. Experimental results on synthetic data and real life images from the remote sensing domain prove the potential of the proposed method. Turning to texture based segmentation, we present a novel multiscale texture model, and a related algorithm for the unsupervised segmentation of color images. Elementary textures are characterized by their spatial interactions with neighboring regions along selected directions. Such interactions are modeled in turn by means of a set of Markov chains, one for each direction, whose parameters are collected in a feature vector that synthetically describes the texture. Based on the feature vectors, the texture are then recursively merged, giving rise to larger and more complex textures, which appear at different scales of observation: accordingly, the model is named Hierarchical Multiple Markov Chain (H-MMC). The Texture Fragmentation and Reconstruction (TFR) algorithm, addresses the unsupervised segmentation problem based on the H-MMC model. The ``fragmentation'' step allows one to find the elementary textures of the model, while the ``reconstruction'' step defines the hierarchical image segmentation based on a probabilistic measure (texture score) which takes into account both region scale and inter-region interactions. The performance of the proposed method was assessed through the Prague segmentation benchmark, based on mosaics of real natural textures, and also tested on real-world natural images from the Berkeley dataset. Finally, an application of the TFR algorithm for the unsupervised segmentation of multiresolution remotely sensed images is proposed, based on the significant presence of textural content in last generation high-resolution satellite images. The obtained results further assess the potential of the proposed method for such real life applications.

In this paper, we present a new unsupervised colour image segmentation algorithm using competitive and morphological concepts. The algorithm is carried out in three processing stages. It starts by an estimation of the density function, followed by a training competitive neural network with a new criterion of resemblance called Mahalanobis distance which detects local maxima of the density function, and ends by the extraction of modal regions using an original method based on the morphological concept. The so detected modes are then used for the classification process. Compared to the K-means clustering or to the clustering approaches based on the different competitive learning schemes, the proposed algorithm has proven, under a number of real and synthetic test images, that it is automatic, has a fast convergence and does not need priori information about the data structure.

In this paper, an unsupervised segmentation method using clustering is presented for color images. We propose to use a neural network based approach to automatic feature selection to achieve adaptive segmentation of color images. With a self-organizing feature map (SOFM), multiple color features can be analyzed, and the useful feature sequence (feature vector) can then be determined. The encoded feature vector is used in the final segmentation using fuzzy clustering. The proposed method has been applied in segmenting different types of color images, and the experimental results show that it outperforms the classical clustering method. Our study shows that the feature encoding approach offers great promise in automating and optimizing the segmentation of color images.

In a great number of fields of computer vision image segmentation plays an essential role as a preliminary step towards further and higher levels of image processing. An original algorithm for unsupervised segmentation of color images is here presented; it resorts to a palletized representation of images, it uses the low spatial frequency content of color which is represented in the CIELUV space and in this space it builds and thresholds hue and chroma histograms finding out the color clusters which allow the image to be segmented. Evidence of the algorithm's effectiveness is reported.

In this paper, a new unsupervised hierarchical approach to textured color images segmentation is proposed. To this end, we have designed a two-step procedure based on a grey-scale Markovian over-segmentation step, followed by a Markovian graph-based clustering algorithm, using a decreasing merging threshold schedule, which aims at progressively merging neighboring regions with similar textural features. This hierarchical segmentation method, using two levels of representation, has been successfully applied on the Berkeley Segmentation Dataset and Benchmark (BSDB, Martin et al., 2001). The experiments reported in this paper demonstrate that the proposed method is efficient in terms of visual evaluation and quantitative performance measures and performs well compared to the best existing state-of-the-art segmentation methods recently proposed in the literature.

We present an unsupervised segmentation algorithm which uses Markov random field models for color textures. These models characterize a texture in terms of spatial interaction within each color plane and interaction between different color planes. The models are used by a segmentation algorithm based on agglomerative hierarchical clustering. At the heart of agglomerative clustering is a stepwise optimal merging process that at each iteration maximizes a global performance functional based on the conditional pseudolikelihood of the image. A test for stopping the clustering is applied based on rapid changes in the pseudolikelihood. We provide experimental results that illustrate the advantages of using color texture models and that demonstrate the performance of the segmentation algorithm on color images of natural scenes. Most of the processing during segmentation is local making the algorithm amenable to high performance parallel implementation.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

A novel color image segmentation method based on finite Gaussian mixture model is proposed in this paper. First, we use EM algorithm to estimate the distribution of input image data and the number of mixture components is automatically determined by MML criterion. Then the segmentation is carried out by clustering each pixel into appropriate component according to maximum likelihood (ML) criterion. The advantage of our method lies in its ability of less relying on initialization and segmenting images in a totally unsupervised manner. Experimental results show that our segmentation method can obtain better results than other methods.

In this paper a new unsupervised nonparametric cooperative adaptive hyperspectral image segmentation approach is presented. In the approach the images are processed band by band, and then the results of all bands are fused, to get the final segmentation result. Two unsupervised nonparametric segmentation methods are cooperated, namely Fuzzy C-means (FCM), and Linde-Buzo-Gray (LBG) to segment each band in the image. The originality of the approach relies firstly on, its adaptation to the types of regions in an image (textured, non-textured), secondly, on the validation approach, which is done in two levels: direct validation, and validation by genetic fusion. The approach is tested on different mono-band and multi-band images and shown its efficiency.

Unsupervised Image Segmentation is one of the central issues in Computer Vision. From the viewpoint of exploratory data analysis, segmentation can be formulated as a clustering problem in which pixels or small image patches are grouped together based on local feature information. In this contribution, parametrical distributional clustering (PDC) is presented as a novel approach to image segmentation. In contrast to noise sensitive point measurements, local distributions of image features provide a statistically robust description of the local image properties. The segmentation technique is formulated as a generative model in the maximum likelihood framework. Moreover, there exists an insightful connection to the novel information theoretic concept of the Information Bottleneck (Tishby et al. [17]), which emphasizes the compromise between efficient coding of an image and preservation of characteristic information in the measured feature distributions.The search for good grouping solutions is posed as an optimization problem, which is solved by deterministic annealing techniques. In order to further increase the computational efficiency of the resulting segmentation algorithm, a multi-scale optimization scheme is developed. Finally, the performance of the novel model is demonstrated by segmentation of color images from the Corel data base.KeywordsImage SegmentationClusteringMaximum LikelihoodInformation Theory

The traditional multi-resolution Markov random field (MRMRF) model uses two-component Markov random field model on each resolution, and requires training data to estimate the necessary model parameters, which is unsuitable for unsupervised image segmentation. Under this circumstance, a new multi-resolution Markov random field model with variable potential for unsupervised texture image segmentation is presented. The new model solves this problem by introducing a variable potential function for multi-level logistic distribution (MLL) model on each scale. Using this method, the new model can automatically estimate model parameters and produce accurate unsupervised segmentation results. The results obtained on synthetic texture images and remote sensing images demonstrate that a better segmentation is achieved by our model than the traditional MRMRF model.

Analysis of hydro‐climatological time series and spatiotemporal dynamics of meteorological variables has become critical in the context of climate change, especially in Southern African countries where rain‐fed agriculture is predominant. In this work, we compared modern unsupervised time series and segmentation approaches and commonly used time series models to analyse rainfall regime changes in the coastal, sub‐humid and semi‐arid regions of Southern Africa. Rainfall regimes change modelling and prediction inform farming strategies especially when choosing measures for mixed crop–livestock farming systems, as farmers can decide to do rainwater harvesting and moisture conservation or supplementary irrigation if water resources are available. The main goal of this study was to predict/identify rainfall cluster trends over time using regression with hidden logistic process (RHLP) or hidden Markov model regression (HMMR) supplemented by autoregressive integrated moving average (ARIMA) and Facebook Prophet models. Historical time series rainfall data was sourced from meteorological services departments for selected site over an average period of 55 years. Commonly used approaches forecasted an upward rainfall trend in the coastal and sub‐humid regions and a declining trend in semi‐arid areas with high variability between and within seasons. For all sites, Ljung‐Box Test Statistics suggested the existence of autocorrelation in rainfall time series data. Prediction capabilities were investigated using the root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) which indicated not much difference between ARIMA and Facebook Prophet models. RHLP and HMMR offered a unique clustering and segmentation approach examining between and within‐season rainfall variability. A maximum of 20 unique rainfall clusters with similar trend characteristics were determined as going beyond this brought non‐significant difference to regime changes. A clear trend was exhibited from 1980 going backwards as compared to recent years signifying how unpredictable is rainfall in Southern Africa. The unsupervised approaches predicted a clear cluster trend in coastal than in sub‐humid and semi‐arid and the performance was assessed using Akaike information criteria and log‐likelihood which showed improvement in prediction power as the number of segmentation clusters approaches 20.

We propose a color-image segmentation algorithm by unsupervised classification of pixels. The originality of the proposed approach consists in iteratively identifying pixel classes by taking into account both the pixel color distributions in several color spaces and their spatial arrangement in the image. In order to overcome the difficult problem of the color space choice, the algorithm selects the color space that is well suited to construct the class at each iteration step. The selection criterion is based on connectedness and color homogeneity measures of pixel subsets. In order to tune the sensitivity of segmentation, we introduce a hierarchical criterion that allows us to segment images with different numbers of regions as human observers do. Experiments carried out on the well-known Berkeley segmentation dataset show that this multicolor space approach succeeds in constructing classes that effectively correspond to regions in the image.

Low-key images widely exist in imaging-based systems such as space telescopes, medical imaging equipment, machine vision systems. Unsupervised low-key image segmentation is an important process for image analysis or digital measurement in these applications. In this paper, a novel active contour model with the probability density function (PDF) of gamma distribution for image segmentation is proposed. The flexible gamma distribution is used to describe both of the heterogeneous foreground and dark background in a low-key image. Besides, an unsupervised curve initialization method is also designed in this paper, which helps to accelerate the convergence speed of curve evolution. The effectiveness of the proposed algorithm is demonstrated through comparison with the CV model. Finally, an industrial application based on proposed approach is described in this paper.

The application of a Markov random field (MRF) state model in an expectation-maximization (EM)-based approach to unsupervised image segmentation is investigated. In the calculation of the marginal distribution of the state field, it is shown that the use of the expected state values for interacting pixel sites in the computation of the MRF energy function may be interpreted as a mean-field approximation. The implications of calculating a self-consistent expectation of the state field are considered. EM convergence criteria are considered, and a criterion based upon divergence is proposed. Experimental results based on synthetic data illustrate the performance advantage of the mean-field approximation and the computational advantage of using self-consistent expectations. >

Deep expectation-maximization network for unsupervised image segmentation and clustering