Papers by Keyword: Texture Feature

Abstract: In this paper, a modified adaptive K-means (MAKM) method is proposed to extract the region of interest (ROI) from the local and public datasets. The local image datasets are collected from Bethezata General Hospital (BGH) and the public datasets are from Mammographic Image Analysis Society (MIAS). The same image number is used for both datasets, 112 are abnormal and 208 are normal. Two texture features (GLCM and Gabor) from ROIs and one CNN based extracted features are considered in the experiment. CNN features are extracted using Inception-V3 pre-trained model after simple preprocessing and cropping. The quality of the features are evaluated individually and by fusing features to one another and five classifiers (SVM, KNN, MLP, RF, and NB) are used to measure the descriptive power of the features using cross-validation. The proposed approach was first evaluated on the local dataset and then applied to the public dataset. The results of the classifiers are measured using accuracy, sensitivity, specificity, kappa, computation time and AUC. The experimental analysis made using GLCM features from the two datasets indicates that GLCM features from BGH dataset outperformed that of MIAS dataset in all five classifiers. However, Gabor features from the two datasets scored the best result with two classifiers (SVM and MLP). For BGH and MIAS, SVM scored an accuracy of 99%, 97.46%, the sensitivity of 99.48%, 96.26% and specificity of 98.16%, 100% respectively. And MLP achieved an accuracy of 97%, 87.64%, the sensitivity of 97.40%, 96.65% and specificity of 96.26%, 75.73% respectively. Relatively maximum performance is achieved for feature fusion between Gabor and CNN based extracted features using MLP classifier. However, KNN, MLP, RF, and NB classifiers achieved almost 100% performance for GLCM texture features and SVM scored an accuracy of 96.88%, the sensitivity of 97.14% and specificity of 96.36%. As compared to other classifiers, NB has scored the least computation time in all experiments.

Abstract: In this paper, a simple yet robust algorithm for texture identification using 1 Dimensional Discrete Fourier Transform (1-D DFT) and Dynamic Time Warping (DTW) is presented with illumination variations. In the first stage, several image processing techniques namely Fuzzy C means (FCM) clustering, edge detection, Otsu thresholding and inverse surface thresholding method are utilized to locate the region of interest (ROI) where defects might exist. Next, the image undergoes the feature extraction process using 1-D DFT and finally, the features are classified using DTW. Several defect images consist of 2 types of defect namely the porosity and crack are experimented and classified using the DTW.

Abstract: Aiming at the difficulties in the segmentation for high-resolution remote multispectral sensing images, this paper proposed a segmentation approach for remote sensing images based on texture features. The algorithm implemented precipitation watershed transform respectively on the texture images obtained by the different characteristics of GLCM, and then superimposed the two segmentation results, finally completing the image segmentation by using a novel regional consolidation method that combined the texture features. The experiments were implemented on the high-resolution ALOS and SPOT 5 remote sensing images respectively. Compared with the traditional watershed segmentation approach based on gradient information, the experimental results showed that the proposed algorithm can accurately locate the edges of objects, effectively overcome the phenomenon of over-segmentation and under-segmentation, with a higher segmentation accuracy and stability.

Abstract: Current image retrieval methods use only one kind of image features, which can not describe image content completely. In this paper, an image retrieval method based on color and texture integration is proposed. Calculation of similarity for the individual features is normalized before color feature integration. In order to improve the precision of image retrieval, a relevance feedback mechanism is invoked. The experiment results show that the proposed method has an excellent retrieval performance.

Abstract: A novel adaptive image feature reduction approach for object tracking using vectorized texture feature is proposed in this paper. Our contributions are three-fold: 1) a statistical discriminative appearance model using texture feature was proposed. 2) Majority of dimensions of the features are removed by judging their errors of the chosen distribution model. The remaining dimensions are most discriminative ones for classification task. The dimension reduction has advantages of reducing the computational cost in classification stage. 3) An adaptive learning rate was proposed to handle drifts caused by long term occlusion. Preliminary experimental results are satisfactory and compared to state-of-the-art object tracking methods.

Abstract: Camshift is an efficient single-feature algorithm based on color-kernel for simple object tracking. However, it’s easy to be interfered by sudden illumination variation or the similar color object in the background. Experiments show that the improved algorithm can track object accurately in dealing with sudden illumination variation and the similar color object in the background, and can meet the requirements of the actual task of object tracking.

Abstract: In order to improve above problems, this paper presents a new patch-based sampling algorithm for synthesizing textures from an input sample texture. We determine the size of block analyzing the effect of the best matched texture to synthesis result, and the result is satisfying. Moreover, the patch-based sampling algorithm remains effective when pixel-based non-parametric sampling algorithms fail to produce good results.

Abstract: Textile, as a kind of universal significant material in the human history, becomes an increasing number of popular among the people. Besides, it has the double needs in the material in the interior decoration. With the rapid development of our domestic economy, people, in China, put forward to more higher standard for the demands of their indoor living, The soft decoration dominated by interior textiles has become the mainstream and main development trend of design field in our society. This article intends to analyze from the level of the texture feature, and from the four aspects to analyze the influence of textile texture in the indoor decoration: textile texture, the color of textile, pattern and cultural of textile, and the emotional of textile, meanwhile, around the texture features of the textile and extend the wide inspiration and thinking to the interior adornment.

Abstract: .In this paper, a video tracking approach based on particle filter is proposed. Texture information is used instead of color. In the proposed approach, gray cooccurrence matrices are used as the texture metric. Experimental results show that the proposed algorithm lead to better result than color feature-based particle filter-based video tracking algorithm and is an effective tool for complicated video tracking application.

Abstract: Feature extraction was a critical stage of image retrieval. For the characterization of the contents of an image could be directly affected by feature extraction. The single feature of images could not fully express images content information, make that the precision of Content-Based Image Retrieval (CBIR) be limited. To overcome the short points, texture information was used in the CBIR. The image color feature and texture feature were comprehensive extracted. And an image retrieval system was developed under MATLAB platform. The new method was simulated and verified by MATLAB.