Unsupervised Perceptual Segmentation of Natural Color Images Using Fuzzy-Based Hierarchical Algorithm

For the favour of data partition, colour image segmentation framework is an ideal choice. This is based on colour and texture features of an image. The main objective of the proposed work consists of a new method for the extraction of colour features that assess the input image in a colour representation. To achieve this, the proposed work used the opponent characteristics of the RGB and LAB colour spaces. The texture features are computed using a gabor filter. The prime role of this work rests in the integration of the colour and texture features with the aim of identifying the homogenous regions in the image. The accuracy for extraction of colour texture was established by the proposed method.

This paper presents a segmentation approach for color natural images that performs an edge-preserving smoothing preprocess and a fuzzy-based region growing segmentation. The hue-chroma-intensity (HCI) color space is adopted for its perceptual uniformity. The edge-preserving smoothing preprocess is performed to smooth unnecessary details of an original image and obtain stable HCI information before transformation of color space and image segmentation. The proposed segmentation algorithm is performed in two processes: pixel-label assignment and region merging, which builds up fuzzy-based homogeneity measures for any two neighboring pixels, any pixel with its adjacent regions and any two adjacent regions by making a soft fusion of HCI information. We make some experiments to verify the necessity of the preprocess and the effectiveness of the proposed segmentation algorithm for rough segmentation of color natural images.

Image segmentation is one of the most attractive problems in image processing. In image segmentation how to extract useful features from image has become crucial. However, color feature or texture feature, which are both wildly used features, could not process segmentation problem alone very well, especially when images are complex. We adopt a rough-fuzzy set approach, which can properly process high dimensionality, for image segmentation considering both color and texture features. This approach firstly constructs a structure named fuzzy data cube, whose attributes are composed of the fuzzy sets associated with image features. The fuzzy data cube, which can be two-dimension or high-dimension, is as the basic data structure in this method. A definition of the membership function of similarity relation based rough-fuzzy set is introduced as well as the definition of dependency function to evaluate the importance of an attribute for image segmentation. Then we used the rough-fuzzy set to discover the similarity set in fuzzy data cube to obtain the segmentation result. Experiments on mosaic and natural images are presented to demonstrate the effectiveness of the proposed method.

This paper presents a color recognition and segmentation of color image through a combination of a multi-scale fuzzy system, an image edge based on Prewitt operator and morphological operators. In the proposed algorithm, after converting image from RGB color space to hue-saturation-value (HSV) color space, different scales of fuzzy sets in hue channel is constructed according to the people's psychological cognition to color. The image color would be quickly recognized fuzzily as less than ten kinds on scale one. To improve the performance of the color recognition rate, recognized image need to be identified on scale two to distinguish different colors observed by human eyes with the same color scheme. Then, adopting Prewitt operator to extract image edge information and applying a sequence of some morphological operators to the edge image, completing the multi-scale segmentation of image with the integration of color and edge information. Finally, indicating its area and species of specific color. The proposed algorithm can automatically adjusted in the process of color recognition and segmentation of image, it achieves recognition and segmentation without supervision. The obtained result has a good consistency with man's subjective visual perception.

Image retrieval is the process of finding images in the database that are similar to the query image by measuring how close the feature values of the query image are to other images. Image retrieval is currently dominated by approaches that combine several different representations or features. The optimal weight of each feature is needed in combining the image features such as color features, texture features, and shape features. In this study, we use a multi-layer perceptron artificial neural network (MLP) method to obtain feature weights automatically and simultaneously look for optimal weights. The color moment is used to find nine color features, Gray Level Co-occurrence Matrix (GLCM) to find four texture features, and Hu Moment to find seven shape features totaling 20 neurons and all of these features become the input layer in our MLP network. Three neurons in output layers become the automatic weight of each feature. These weights are used to combine each feature's role in obtaining the relevant image. Euclidean Distance is used to measure similarity. The average precision values obtained using automatic feature weights are 93.94% for the synthetic dataset, 91.19% for the Coil-100 dataset, and 54.31% for the Wang dataset. These results have an average difference of 5.06% with the target so automatic feature weighting works well. This value is obtained at a hidden layer size of 11 and a learning rate of 0.1. In addition, the use of automatic feature weighting gives more accurate results compared to manual feature weighting.

In this study, an identification model based on computer vision and artificial neural network technologies is proposed for the identification of the ripeness of fresh corn ears. For collected images of corn ears, 2D discrete wavelet transform is performed to extract information of low-frequency sub-band as color features, and discrete Fourier transform is performed to extract energy spectrum information as texture features. Principle component analysis is employed for the fusion and dimensionality reduction of color and texture features, and the first three principle components are chosen as inputs of the network model in order to establish probabilistic neural network model for the automated ripeness identification of fresh corn ears. Simulation analysis demonstrates that the identification accuracy of this model reaches 90.67%.

Color texture segmentation based on image pixel classification

The segmentation of acne vulgaris lesions is a crucial step for the classification and developing a severity assessment system. In this paper, an unsupervised technique is proposed for the segmentation of acne vulgaris lesions. The effect of color feature normalization on segmentation results is examined. Two types of normalization is performed — in the first type, the color feature values are restricted to the range [0 1] and in the second type, the color features are normalized such that their mean value is zero and their variance is one. The segmentation of acne lesions was carried out in the feature set formed by combining chromatic components from different color spaces such as YIQ, Lab, I1I2I3 and YCbCr. A dataset of fifty color images of acne patients have been used in this experimentation. The segmentation results show that normalization can significantly improve the segmentation results in YIQ and I1I2I3 color spaces. It has also found that comparatively good segmentation results can be obtained in high dimensional chromatic space.

ファジィ分離統合法によるカラー自然画像の大まかな領域分割とダーモスコピー画像への応用(ITS画像処理,映像メディア及び一般)

Building extraction is an important way to obtain information in urban planning, land management, and other fields. As remote sensing has various advantages such as large coverage and real-time capability, it becomes an essential approach for building extraction. Among various remote sensing technologies, the capability of providing 3D features makes the LiDAR point cloud become a crucial means for building extraction. However, the LiDAR point cloud has difficulty distinguishing objects with similar heights, in which case texture features are able to extract different objects in a 2D image. In this paper, a building extraction method based on the fusion of point cloud and texture features is proposed, and the texture features are extracted by using an elevation map that expresses the height of each point. The experimental results show that the proposed method obtains better extraction results than that of other texture feature extraction methods and ENVI software in all experimental areas, and the extraction accuracy is always higher than 87%, which is satisfactory for some practical work.

Skin color is a robust cue in human skin detection. It has been widely used in various human-related image processing applications. Although many researches have been carried out for skin color detection, there is no consensus on which color space is the most appropriate for skin color detection because many researchers do not provide strict justification of their color space choice. In this paper, a comprehensive comparative study using the Multilayer Perceptron artificial neural network (MLP), which is a universal classifier, is carried out to evaluate the overall performance of different color-spaces for skin detection. It aims at determining the most optimal color space using color and color-texture features separately. The study has been carried out using images of different databases. The experimental results showed that the YIQ color space gives the highest separability between skin and non-skin pixels among the different color spaces tested using color features. Combining color and texture eliminates the differences between color spaces but leads to much more accurate and efficient skin detection.

Radiomics has shown potential in disease diagnosis, but its feasibility for non-small cell lung carcinoma (NSCLC) subtype classification is unclear. This study aims to explore the diagnosis value of texture and colour features from positron emission tomography computed tomography (PET-CT) images in differentiation of NSCLC subtypes: adenocarcinoma (ADC) and squamous cell carcinoma (SqCC). Two patient cohorts were retrospectively collected into a dataset of 341 18F-labeled 2-deoxy-2fluoro-d-glucose ([18F] FDG) PET-CT images of NSCLC tumours (125 ADC, 174 SqCC, and 42 cases with unknown subtype). Quantification of texture and colour features was performed using freehand regions of interest. The relation between extracted features and commonly used parameters such as age, gender, tumour size, and standard uptake value (SUVmax) was explored. To classify NSCLC subtypes, support vector machine algorithm was applied on these features and the classification performance was evaluated by receiver operating characteristic curve analysis. There was a significant difference between ADC and SqCC subtypes in texture and colour features (P < 0.05); this showed that imaging features were significantly correlated to both SUVmax and tumour diameter (P < 0.05). When evaluating classification performance, features combining texture and colour showed an AUC of 0.89 (95% CI, 0.78–1.00), colour features showed an AUC of 0.85 (95% CI, 0.71–0.99), and texture features showed an AUC of 0.68 (95% CI, 0.48–0.88). DeLong’s test showed that AUC was higher for features combining texture and colour than that for texture features only (P = 0.010), but not significantly different from that for colour features only (P = 0.328). HSV colour features showed a similar performance to RGB colour features (P = 0.473). The colour features are promising in the refinement of NSCLC subtype differentiation, and features combining texture and colour of PET-CT images could result in better classification performance.

We present an efficient approach to index and retrieve Roman mosaic-images by region. The image regions are defined based on the mosaic-image presentation. For more efficiency, the proposed approach uses an index combining texture and fuzzy colour features. These features are fuzzy image dominant colour, fuzzy colour histogram and local Robert gradient binary pattern operator LRGBP. Similar images are extracted using a fuzzy similarity and histogram intersection measures. Our proposed index is tested on database containing 800 images of historical Roman mosaics from the 1st to 4th century. Tests demonstrate that the fusion of the fuzzy colour and texture features increases the precision of the retrieval system. Evaluation tests and comparison with several other prevailing approaches prove a remarkable performance of our proposed approach.

This article introduces a novel class of active contour models for image segmentation. It makes use of nonlocal comparisons between pairs of patches within each region to be segmented. The corresponding variational segmentation problem is implemented using a level set formulation that can handle an arbitrary number of regions. The pairwise interaction of features constrains only the local homogeneity of image features, which is crucial in capturing regions with smoothly spatially varying features. This segmentation method is generic and can be adapted to various segmentation problems by designing an appropriate metric between patches. We instantiate this framework using several classes of features and metrics. Piecewise smooth grayscale and color images are handled using $L^2$ distance between image patches. We show examples of efficient segmentation of natural color images. Locally oriented textures are segmented using the $L^2$ distance between patches of Gabor coefficients. We use a Wasserstein distance between local empirical distributions for locally homogeneous random textures. A correlation metric between local motion signatures is able to segment piecewise smooth optical flows.

Image segmentation is critical to image processing and pattern recognition. Basically, color image segmentation techniques are based on monochrome ones operating in different color spaces. A color image segmentation method in the RGB color space is reported in the paper, the image segmentation is used in the image about the disease spot and the normal spot of cucumber in the greenhouse, we determine the RGB threshold value, through counting the characteristic parameter of stochastic picture element in the different image spot. The combined methods of the main factor three threshold value R, G B decide the division. this method not only retain the image real color, but also have very good division effect.