Color Image Denoising Using Gaussian Multiscale Multivariate Image Analysis

In the real world, the biological visual system is more efficient than the machine visual system in analyzing visual information. Physiology theories show that this efficiency owes to the multi-layer neural network in human visual system, in which every layer accomplishes different tasks and is related with other layers. The low-level stages of the human visual system, especially the retina, can provide certain scale information for the high-level stages of visual system through using the non-classical receptive field (nCRF) mechanism. This mechanism that the nCRF size can be adjusted automatically by ganglion cell (GC) can achieve a multi-scale image analysis. The results, reflecting the distribution of the image information, can be shared by several algorithms or processes solving different visual tasks, such as contour detection and image segmentation. A model of multi-scale image analysis based on GC has been proposed in this paper, which retains the key information and reduces the redundancy information for the further stages of the visual system. Experimental results on N-cut and contour detection show that this multi-scale image analysis model provides distinctive improvement for these image processing tasks.

Urban is composed of different critical levels of organization where interactions are stronger within levels than among levels, it is a complex system. Image analysis performed at a unique scale for urban applications is doomed to be incomplete and misleading. This paper presents a multiscale image analysis method based on wavelet transform and watershed transform for urban system. It performs image analysis on multiple scale images, and can catch comprehensive information of urban system for different applications. The procedure consists of three main steps: multiscale representation, image segmentation, post-processing. First, multiscale represent creates multiresolution images using a wavelet transform and a modified inverse wavelet transform; a smooth and multiscale presentation of original image is provided. Second, image segmentation segments the multiscale images using watershed segment algorithm. Finally, post-processing merges over-segment regions and smoothes watersheds, then, multiscale image objects are attained. This study was conducted in Nanjing, Jiangsu province, China, using IKONOS satellite data. Experimental results demonstrate that the proposed method is suitable for describing the complex urban system. Both visual and quantitative methods are applied to evaluation the experimental results, and reveal the relation of objects within and among levels.

Colorectal cancer (CRC) is markedly heterogeneous and develops progressively toward malignancy through several stages which include stroma (ST), benign hyperplasia (BH), intraepithelial neoplasia (IN) or precursor cancerous lesion, and carcinoma (CA). Identification of the malignancy stage of CRC pathology tissues (PT) allows the most appropriate therapeutic intervention. This study investigates multiscale texture features extracted from CRC pathology sections using 3D wavelet transform (3D-WT) filter. Multiscale features were extracted from digital whole slide images of 39 patients that were segmented in a pre-processing step using an active contour model. The capacity for multiscale texture to compare and classify between PTs was investigated using ANOVA significance test and random forest classifier models, respectively. 12 significant features derived from the multiscale texture (i.e., variance, entropy, and energy) were found to discriminate between CRC grades at a significance value of p < 0.01 after correction. Combining multiscale texture features lead to a better predictive capacity compared to prediction models based on individual scale features with an average (±SD) classification accuracy of 93.33 (±3.52)%, sensitivity of 88.33 (± 4.12)%, and specificity of 96.89 (± 3.88)%. Entropy was found to be the best classifier feature across all the PT grades with an average of the area under the curve (AUC) value of 91.17, 94.21, 97.70, 100% for ST, BH, IN, and CA, respectively. Our results suggest that multiscale texture features based on 3D-WT are sensitive enough to discriminate between CRC grades with the entropy feature, the best predictor of pathology grade.

To address the challenges of manual inspection dependency, low efficiency, and high costs in evaluating the surface grinding quality of composite materials, this study investigated machine vision-based surface recognition algorithms. We proposed a multi-scale texture fusion analysis algorithm that innovatively integrated luminance analysis with multi-scale texture features through decision-level fusion. Specifically, a modified Rayleigh parameter was developed during luminance analysis to rapidly pre-segment unpolished areas by characterizing surface reflection properties. Furthermore, we enhanced the traditional Otsu algorithm by incorporating global grayscale mean (μ) and standard deviation (σ), overcoming its inherent limitations of exclusive reliance on grayscale histograms and lack of multimodal feature integration. This optimization enables simultaneous detection of specular reflection defects and texture uniformity variations. To improve detection window adaptability across heterogeneous surface regions, we designed a multi-scale texture analysis framework operating at multiple resolutions. Through decision-level fusion of luminance analysis and multi-scale texture evaluation, the proposed algorithm achieved 96% recognition accuracy with >95% reliability, demonstrating robust performance for automated surface grinding quality assessment of composite materials.

To help improve efficacy of screening mammography by eventually establishing a new optimal personalized screening paradigm, the authors investigated the potential of using the quantitative multiscale texture and density feature analysis of digital mammograms to predict near-term breast cancer risk. The authors' dataset includes digital mammograms acquired from 340 women. Among them, 141 were positive and 199 were negative/benign cases. The negative digital mammograms acquired from the "prior" screening examinations were used in the study. Based on the intensity value distributions, five subregions at different scales were extracted from each mammogram. Five groups of features, including density and texture features, were developed and calculated on every one of the subregions. Sequential forward floating selection was used to search for the effective combinations. Using the selected features, a support vector machine (SVM) was optimized using a tenfold validation method to predict the risk of each woman having image-detectable cancer in the next sequential mammography screening. The area under the receiver operating characteristic curve (AUC) was used as the performance assessment index. From a total number of 765 features computed from multiscale subregions, an optimal feature set of 12 features was selected. Applying this feature set, a SVM classifier yielded performance of AUC = 0.729 ± 0.021. The positive predictive value was 0.657 (92 of 140) and the negative predictive value was 0.755 (151 of 200). The study results demonstrated a moderately high positive association between risk prediction scores generated by the quantitative multiscale mammographic image feature analysis and the actual risk of a woman having an image-detectable breast cancer in the next subsequent examinations.

Reliable and accurate land use/land cover (LULC) map is a crucial data source for the understanding of coupled human–environment systems, monitoring changes, timely low-cost planning, and management of natural resources. Improvements in sensor technologies and machine learning capabilities have shifted the attention of remote sensing community to data complementarity through fusion of multi-sensor data for accurate feature extraction and mapping. Amalgamation of optical and synthetic aperture radar (SAR) images has shown promising advantages in enhancing the accuracy of extracting LULC as such method allows exploitation of information in sensors. This study investigated the potential of using freely available multisource Sentinel images to extract LULC maps in semi-arid environments through multi-scale geographic object-based image analysis (GEOBIA). A multi-scale classification framework that integrates GEOBIA, correlation-based feature selection (CFS), and random forest (RF)-supervised classification was adopted to extract LULC from assimilation of Sentinel multi-sensor products. First, Sentinel-1 and -2 images were pre-processed. Second, optimum multi-scale segmentation levels were selected using F-score segmentation quality measures. Third, 70 features of various spectral indices and derivatives and geometrical features from optical data and multiple ratios and textural features from dual-polarization SAR images were computed, and a CFS based on wrapper approach was used to select the most significant features at multi-scale levels. Finally, a single and multi-scale RF classifier was used to extract LULC classes using the most relevant features extracted from Sentinel SAR and optical images. Results of multi-scale image segmentation optimization showed that scale parameter (SP) values of 40, 60, and 150 were optimal for extraction of LULC classes. Results of feature selection showed that 22, 24, and 27 features were selected at scale SP values of 40, 60, and 150, respectively. Half of the features were common among the three scales. Single RF classification yielded overall accuracy (OA) values of 92.10%, 93%, and 91% and kappa coefficients of 0.901, 0.912, and 0.89 at scale values of 150, 60, and 40, respectively. Multiscale RF classification from scale values of 150 and 60 produced better LULC classification with OA 96.06% and kappa coefficient of 0.95 compared with other scale SP values. The integrated approach demonstrated an effective and promising method for high-quality LULC extraction from coupling optical and SAR images. Overall, multi-sensor Sentinel images along with the adopted approach feature a remarkable potential for improving LULC extraction and can effectively be used to update geographic information system layers for various applications.

In this paper, a new multiscale texture segmentation method using contextual hidden Markov tree (CHMT) in wavelet domain is proposed. The hidden Markov tree models (HMT) describe persistence property of wavelet coefficients in multiscale images, but loses clustering property. The method is put forward to overcome the shortcoming of standard HMT by using extended coefficients without changing the wavelet tree structure and makes it possible to get a more accurate segmentation result. Experimental results demonstrate that the proposed method is effective for multiscale texture image segmentation.

Quadtree representation presents many advantages in image analysis because of its better manipulation at different scales. In computer vision one frequently faces the problem of multi-scale analysis. We use a syntactic representation of quadtree to analyse images in the Haar base with pyramidal algorithms in linear time. We present applications in dynamic filtering, in compression and image segmentation.

Towards intelligent robust detection of anatomical structures in incomplete volumetric data.

Some special topics in multivariate image analysis

A simple imaging system has been developed for acquiring multivariate images in order to characterise the heterogeneity of food materials. The objective of the present work is, first, to demonstrate the capability of this acquisition system to discriminate food products of different natures. Secondly, our goal is to apply Partial Least Squares regression on these multivariate images and to evaluate the interest of various strategies of classification. A data set containing 24 images (702 × 524) acquired at different wavelengths for four food products is analysed. After the establishment of the PLS2 models employed for predicting the indicator variables, four strategies of classification of observations are tested. The first classification is done by selecting the largest component of the indicator variables. The others are based on the measurement of distances to the barycentres of the qualitative groups. Distances calculated can be either Euclidian distances or Mahalonobis distances. Except the strategy based on the Euclidian distance on scores, the strategies are rather equivalent, with a slight advantage to the Euclidian distance on predicted indicators. Another possibility addressed by the use of linear discriminant analysis (LDA) on multivariate images is to represent the qualitative groups as artificial images. The largest confusion appears between both cereal products while others are well classified. Copyright © 2006 John Wiley &amp; Sons, Ltd.

In this work we improve a methodology for discriminating urban environments by means of textural features in SAR images. In particular, we introduce multi-scale co-occurrence features and show how the feature set may be chosen as a function of the training set and the mapping classes. Moreover, we provide and compare results obtained by different satellite SAR sensors on the same urban test site, as well as a combination of these sets. Finally, a short analysis of the polarization effects and their importance in this framework of analysis is considered. The results are extremely encouraging, and show the potential of this technique, even if more research is needed to exploit the capabilities of the new generation of low-Earth orbit SAR satellites.

This paper introduces new features for describing image patterns. We integrate the concepts of multiscale image analysis, aura matrix to define image features, and to obtain the features having robustness with illumination variations and shading effects, we analyse images based on the topographic structure described by the surface-shape operator. Then, illustrate usefulness of the proposed features with texture classifications. Results show that the proposed features extracted from multiscale images work much better than those from a single scale image, and confirm that the proposed features have robustness with illumination and shading variations. By comparisons with the multiresolution simultaneous autoregressive features using Mahalanobis distance and Euclidean distance, the proposed multiscale features give better performances for classifying the entire Brodatz textures (1966).

Statistic of natural images is a growing field of research both in vision and image processing. On the vision research side, fine statistical details about object distribution in real-world scenes help understanding the human visual system behavior. On the image processing side, by using the information gathered from statistics of natural scenes, we can obtain reliable priors and insights that can be used in many models. In has been rigorously proven in [ 16 ] that, if second order stationarity and commutativity of spatiochromatic covariance matrices hold true for natural scenes, then the codification of spatial and chromatic information by the human visual system can be separated through a tensor product. Spatial features are coded via local and oriented Fourier basis elements, while color features are coded via a triad given by an achromatic channel followed by two color opponent channels. In this paper, we will show that, while stationarity is guaranteed, commutativity is not. However, we shall see that commutativity of spatiochromatic covariance matrices can be approached if the database of images used to model visual scenes is modified accordingly to a suitable transformation that describes the response of retinal photoreceptors to light absorption: the Michaelis-Menten formula. A thorough investigation of the effects of a parameter of this formula will be performed and its influence on commutativity of covariance matrices will be detailed.

Multi-scale texture loss for CT denoising with GANs