Haralick Features Research Articles

Lung nodules classification in CT images using texture descriptors

In lung cancer, early diagnosis can improve potentially the prognosis. Accurate interpretation of computed tomography (CT) scans demands significant efforts by radiologists due to the extensive number of slices analyzed in each examination, for each patient. Computer-aided diagnosis (CAD) systems have been applied in several medical fields, but mostly in lung nodules detection and classification. CAD systems for lung lesions classification usually extract different types of features from lesions, such as texture feature, shape and intensity. This exploratory study aims to investigate the performance of lung nodules classification in 2D and 3D CT lesions images using Haralick texture features analysis and binary logistic regression. Expert radiologists manually segmented from a CT dataset of 17 benign and 20 malignant nodules, which have their anatomopathological results. Haralick features were extracted from 2D lesions images, using the largest cross-section nodule area, and from all nodule volume (3D). Principal Component Analysis (PCA) was applied to reduce texture features dimensionality, showing two and three principal components (PC) can explain 85.8% and 96.25% of data variance for 2D lesions, and 72.4% and 91.7% for 3D lesions, respectively. Binary logistic regression using leave-one-out cross-validation for training and test datasets showed no differences in accuracy (63% - 68%), using two or three PC. The higher sensitivity (75%) was acquired using 2D images with two or three PC, while the higher specificity (65%) was obtained using 3D images with two or three PC. Binary logistic regression using a small number of Haralick texture features showed better accuracy in lung nodules classification than visual evaluation by radiologists, although the limited dataset. Further studies are needed to generalize and improve these results.

Cryptographic Algorithm based Feature Level Fusion of Fingerprint and Iris in a Multi-Biometric Recognition System

Biometric encryption is one of the developing exploration area, which is a strategy for merging biometric features with cryptographic keys. Biometric Recognition is based on the anatomical and behavior attributes of the individuals. Multibiometric is the combination of various biometrics like Fingerprint, Iris, and Face, Fingervein etc. Experts are concentrating on the most proficient method to give security to the framework, the template which was produced from the biometric should be ensured. The main objective of this paper is to protect the multi biometric template by creating a protected sketch by deploying bio cryptosystem. Once the biometric template is stolen it turns into a major issue for the security of the framework and furthermore for client protection. In this way, a bio-crypto framework ensures the confidentiality of the information. In this paper bio cryptosystem is proposed to improve the security of multimodal frameworks by producing the biocrypto key from Finger print and iris. Gray level co-occurrence matrix (GLCM) based Haralick features, local binary pattern (LBP), triplet half-band filter bank (THFB) and dynamic features (DF) are extracted from fingerprint and iris. The high dimensionality space of the features are reduced using kernel principal component analysis (KPCA. Finally, the encoding process is matted with biometric key utilizing symmetric RSA (Rivest-Shamir-Adleman) cryptographic algorithm.

Analysis of internet of things malware using image texture features and machine learning techniques

Threats emanating from Internet of Things (IoT) malicious software are fast evolving and complex phenomena. Unlike conventional networks, Internet of things have unique attributes such as heterogeneity of devices, high scalability and diverse architectures that makes its malware analysis complex. In this paper, we propose an approach for analyzing and classifying IoT malware using Haralick image texture features and machine learning classifiers namely K-nearest neighbor(KNN), naïve Bayes (NB) and random forest(RF). A binary file (malicious or benign) is converted to a gray scale image. The gray level co-occurence matrix (GLCM) is computed on each of the extracted image. On the basis of these GLCM parameters, five Haralick features namely angular second moment, entropy, contrast, inverse different moment and correlation are calculated. Finally, these Haralick texture features are used to perform malware classification using random forest, naïve Bayes and K-nearest neighbor. Experimental results shows that our approach obtains 95% accuracy for Random Forest, 89% for naïve Bayes and 80% for K-nearest neighbor classifiers. Overall, use of texture feature can realize a low computational and platform independent classification scheme for IoT malware.

PUG-FB : Person-verification using geometric and Haralick features of footprint biometric

This article demonstrates a study of biometric identification and verification system using foot geometry features. A footprint has three types of features which are sufficient to recognize a person uniquely. These features are categorized into geometric, texture, and minutiae. We have computed most widely used geometry features of the foot using length, width, area, major axis, and minor axis, to identify a person uniquely. Different variations of these features are computed by assigning weights to each feature emphasizing its importance. We have extracted the best variations among foot descriptors, and conclude that the province is the most contributing factor to identify a person foot uniquely. Foot contour features are further combined with foot descriptors to increase the accuracy. For texture, Gray level co-occurrence matrix based on Haralick features is computed with Support Vector Machine as the classifier. Foot biometrics can be used as an additional covert authentication measure where people remove shoes, such as holy places, airport security, swimming pools, wellness centers etc. It can also be used for newborn authentication and identification in hospitals. The method achieves GenuineAcceptRate(GAR) of 82% with the FalseAcceptRate(FAR) of 2.0%, and GAR of 85% with the FAR of 4.0% in case of combination sum rule. GenuineAcceptRate(GAR) has increased to 87.5% at FalseAcceptRate(FAR) of 2.0% including texture features as Gray level co-occurrence matrix.

Mangrove Forest Classification in Drone Images Using HSV Color Moment and Haralick Features Extraction with K-Nearest Neighbor

Due to the problems with uncontrolled changes in mangrove forests, a forest function management and supervision is required. The form of mangrove forest management carried out in this study is to measure the area of mangrove forests by observing the forests using drones or crewless aircraft. Drones are used to take photos because they can capture vast mangrove forests with high resolution. The drone was flown over above the mangrove forest and took several photos. The method used in this study is extracting color features using mean values, standard deviations, and skewness in the HSV color space and texture feature extraction with Haralick features. The classification method used is the k-nearest neighbor method. This study conducted three tests, namely testing the accuracy of the system, testing the distance method used in the k-nearest neighbor classification method, and testing the k value. Based on the results of the three tests above, three conclusions obtained. The first conclusion is that the classification system produces an accuracy of 84%. The second conclusion is that the distance method used in the k-nearest neighbor classification method influences the accuracy of the system. The distance method that produces the highest accuracy is the Euclidean distance method with an accuracy of 84%. The third conclusion is that the k value used in the k-nearest neighbor classification method influences the accuracy of the system. The k-value that produces the highest accuracy is k = 3, with an accuracy of 84%.

Energy enhanced tissue texture in spectral computed tomography for lesion classification

Tissue texture reflects the spatial distribution of contrasts of image voxel gray levels, i.e., the tissue heterogeneity, and has been recognized as important biomarkers in various clinical tasks. Spectral computed tomography (CT) is believed to be able to enrich tissue texture by providing different voxel contrast images using different X-ray energies. Therefore, this paper aims to address two related issues for clinical usage of spectral CT, especially the photon counting CT (PCCT): (1) texture enhancement by spectral CT image reconstruction, and (2) spectral energy enriched tissue texture for improved lesion classification. For issue (1), we recently proposed a tissue-specific texture prior in addition to low rank prior for the individual energy-channel low-count image reconstruction problems in PCCT under the Bayesian theory. Reconstruction results showed the proposed method outperforms existing methods of total variation (TV), low-rank TV and tensor dictionary learning in terms of not only preserving texture features but also suppressing image noise. For issue (2), this paper will investigate three models to incorporate the enriched texture by PCCT in accordance with three types of inputs: one is the spectral images, another is the co-occurrence matrices (CMs) extracted from the spectral images, and the third one is the Haralick features (HF) extracted from the CMs. Studies were performed on simulated photon counting data by introducing attenuation-energy response curve to the traditional CT images from energy integration detectors. Classification results showed the spectral CT enriched texture model can improve the area under the receiver operating characteristic curve (AUC) score by 7.3%, 0.42% and 3.0% for the spectral images, CMs and HFs respectively on the five-energy spectral data over the original single energy data only. The CM- and HF-inputs can achieve the best AUC of 0.934 and 0.927. This texture themed study shows the insight that incorporating clinical important prior information, e.g., tissue texture in this paper, into the medical imaging, such as the upstream image reconstruction, the downstream diagnosis, and so on, can benefit the clinical tasks.

Multi‐scale analysis of ulcer disease detection from WCE images

Wireless capsule endoscopy (WCE) proves its robustness as a great technology to examine the entire digestive tract or the small intestine. An automatic computer-aided design method is proposed in this study, in a manner to differentiate between ulcer disease and normal WCE images. A multi-scale analysis-based grey-level co-occurrence matrix (GLCM) is conducted here. The main step, the co-occurrence matrix (GLCM), is computed from each sub-band Laplacian pyramid decomposition, so as to extract the common Haralick features. Moreover, the p-value and area under the curve are used to select the relevant characteristics from the feature descriptor. This proposed approach was separately applied to the components of CIELab colour space. Ulcer detection was performed using the support vector machine. The findings demonstrate an encouraging detection rate performance of 95.38% for accuracy and 97.42% for sensitivity based on the first dataset and an average accuracy of 99.25 and 98.51% of sensitivities for the second dataset.

HEMIGEN: Human Embryo Image Generator Based on Generative Adversarial Networks.

We propose a method for generating the synthetic images of human embryo cells that could later be used for classification, analysis, and training, thus resulting in the creation of new synthetic image datasets for research areas lacking real-world data. Our focus was not only to generate the generic image of a cell such, but to make sure that it has all necessary attributes of a real cell image to provide a fully realistic synthetic version. We use human embryo images obtained during cell development processes for training a deep neural network (DNN). The proposed algorithm used generative adversarial network (GAN) to generate one-, two-, and four-cell stage images. We achieved a misclassification rate of 12.3% for the generated images, while the expert evaluation showed the true recognition rate (TRR) of 80.00% (for four-cell images), 86.8% (for two-cell images), and 96.2% (for one-cell images). Texture-based comparison using the Haralick features showed that there is no statistically (using the Student’s t-test) significant (p < 0.01) differences between the real and synthetic embryo images except for the sum of variance (for one-cell and four-cell images), and variance and sum of average (for two-cell images) features. The obtained synthetic images can be later adapted to facilitate the development, training, and evaluation of new algorithms for embryo image processing tasks.

Multisite evaluation of radiomic feature reproducibility and discriminability for identifying peripheral zone prostate tumors on MRI.

Recent advances in the field of radiomics have enabled the development of a number of prognostic and predictive imaging-based tools for a variety of diseases. However, wider clinical adoption of these tools is contingent on their generalizability across multiple sites and scanners. This may be particularly relevant in the context of radiomic features derived from T1- or T2-weighted magnetic resonance images (MRIs), where signal intensity values are known to lack tissue-specific meaning and vary based on differing acquisition protocols between institutions. We present the first empirical study of benchmarking five different radiomic feature families in terms of both reproducibility and discriminability in a multisite setting, specifically, for identifying prostate tumors in the peripheral zone on MRI. Our cohort comprised 147 patient T2-weighted MRI datasets from four different sites, all of which are first preprocessed to correct for acquisition-related artifacts such as bias field, differing voxel resolutions, and intensity drift (nonstandardness). About 406 three-dimensional voxel-wise radiomic features from five different families (gray, Haralick, gradient, Laws, and Gabor) were evaluated in a cross-site setting to determine (a)how reproducible they are within a relatively homogeneous nontumor tissue region and (b) how well they could discriminate tumor regions from nontumor regions. Our results demonstrate that a majority of the popular Haralick features are reproducible in over 99% of all cross-site comparisons, as well as achieve excellent cross-site discriminability (classification accuracy of ). By contrast, a majority of Laws features are highly variable across sites (reproducible in of all cross-site comparisons) as well as resulting in low cross-site classifier accuracies ( ), likely due to a large number of noisy filter responses that can be extracted. These trends suggest that only a subset of radiomic features and associated parameters may be both reproducible and discriminable enough for use within machine learning classifier schemes.

Hybrid computational intelligence algorithms and their applications to detect food quality

Food security is a major problem faced today. With primitive storage facilities, especially in developing countries, it often leads to extensive losses. This work aims to develop algorithms based on vision data to assess the food quality and deploy them in food storage facilities to detect early signs of spoilage. This paper presents various segmentation techniques for finding spoilt food. Novel optimization techniques have been developed and implemented to improve K-means clustering and multilevel thresholding. A hybrid of moth flame optimization (MFO) and gravitational search algorithm (GSA) has been developed. Also, in another hybrid, particle swarm optimization (PSO) was also incorporated along with MFO and GSA. Both the hybrids performed better than the individual algorithms and the MFO–GSA–PSO hybrid performed better than the MFO–GSA hybrid on the benchmark functions. Segmented images using optimized K-means were used for feature extraction using local binary patterns (LBP). Multiclass support vector machine was used for classification which gave an accuracy of 81% for features from segmented images obtained using MFO–GSA hybrid and 83.33% for that using MFO–GSA–PSO hybrid. Results of simple linear iterative clustering superpixels for segmentation have also been discussed. The segmented clusters are then used to judge the rottenness of the food. Classification using LBP and Haralick features of the segmented image obtained using graphs over superpixels gave an accuracy of 81.7% and 78% respectively.

Opportunistic osteoporosis screening in multi-detector CT images via local classification of textures

SummaryOur study proposed an automatic pipeline for opportunistic osteoporosis screening using 3D texture features and regional vBMD using multi-detector CT images. A combination of different local and global texture features outperformed the global vBMD and showed high discriminative power to identify patients with vertebral fractures.IntroductionMany patients at risk for osteoporosis undergo computed tomography (CT) scans, usable for opportunistic (non-dedicated) screening. We compared the performance of global volumetric bone mineral density (vBMD) with a random forest classifier based on regional vBMD and 3D texture features to separate patients with and without osteoporotic fractures.MethodsIn total, 154 patients (mean age 64 ± 8.5, male; n = 103) were included in this retrospective single-center analysis, who underwent contrast-enhanced CT for other reasons than osteoporosis screening. Patients were dichotomized regarding prevalent vertebral osteoporotic fractures (noFX, n = 101; FX, n = 53). Vertebral bodies were automatically segmented, and trabecular vBMD was calculated with a dedicated phantom. For 3D texture analysis, we extracted gray-level co-occurrence matrix Haralick features (HAR), histogram of gradients (HoG), local binary patterns (LBP), and wavelets (WL). Fractured vertebrae were excluded for texture-feature and vBMD data extraction. The performance to identify patients with prevalent osteoporotic vertebral fractures was evaluated in a fourfold cross-validation.ResultsThe random forest classifier showed a high discriminatory power (AUC = 0.88). Parameters of all vertebral levels significantly contributed to this classification. Importantly, the AUC of the proposed algorithm was significantly higher than that of volumetric global BMD alone (AUC = 0.64).ConclusionThe presented classifier combining 3D texture features and regional vBMD including the complete thoracolumbar spine showed high discriminatory power to identify patients with vertebral fractures and had a better diagnostic performance than vBMD alone.

The role of radiomics in the preoperative diagnosis of renal cell carcinoma with sarcomatoid dedifferentiation (sRCC).

621 Background: sRCC is an aggressive renal malignancy, with poor survival and limited response to therapy. Preoperative identification of sRCC would be helpful for counselling patients, and clinical trial enrollment. This study aims at assessing the potential of radiomics to discriminate clear cell sRCC from non-sarcomatoid clear cell RCC (nsRCC). Methods: The study included 49 sRCC and 41 nsRCC patients treated with surgery between 2007-2016, who had contrast-enhanced CT available. An experienced radiologist delineated the entire tumor using 3D Slicer ( http://www.slicer.org ). The extracted 3D region of interest was imported in our in-house radiomic pipeline. A total of 310 features (10 histogram-based and 300 second-order features) were calculated. Second-order radiomic features were calculated using the Grey Level Cooccurrence Matrix (GLCM) and 20 Haralick features were obtained from the GLCM. To account for directionality, the mean, variance and range of the features across different directions were calculated. Finally, different number of gray levels were also considered in the analysis (N = 8, 16, 32, 64, 256). Core features were obtained using a feature selection based on Least Absolute Shrinkage and Selection Operator (LASSO). Selected features were used to build a classification model for prediction of sRCC versus nsRCC (XGboost). To evaluate the robustness of the estimates, Leave One Out Cross-Validation (LOOCV) was conducted on the patient set. Results: Overall, median tumor size was 10.0 cm and most patients had pT3a (68%). There was no significant difference of age, gender, race, tumor size and stages between sRCC and nsRCC cohorts. The prediction of sRCC using LOOCV was significant with p-value &lt; 0.0001. Area under the curve, sensitivity, and specificity for identification of sRCC were 96.8%, 92.6% and 93.8% respectively. Conclusions: This study demonstrates that CT radiomic features can accurately discriminate between sRCC and nsRCC. The proposed tool has the potential to advance clinical management strategies. In addition to being noninvasive, this methodology can be applied to scans obtained during routine clinical care. Further external validation is warranted.

Gray-level invariant Haralick texture features.

Haralick texture features are common texture descriptors in image analysis. To compute the Haralick features, the image gray-levels are reduced, a process called quantization. The resulting features depend heavily on the quantization step, so Haralick features are not reproducible unless the same quantization is performed. The aim of this work was to develop Haralick features that are invariant to the number of quantization gray-levels. By redefining the gray-level co-occurrence matrix (GLCM) as a discretized probability density function, it becomes asymptotically invariant to the quantization. The invariant and original features were compared using logistic regression classification to separate two classes based on the texture features. Classifiers trained on the invariant features showed higher accuracies, and had similar performance when training and test images had very different quantizations. In conclusion, using the invariant Haralick features, an image pattern will give the same texture feature values independent of image quantization.

A novel high-performance holistic descriptor for face retrieval

Texture extraction-based classification has become the facto methodology applied in face recognition. Haralick feature extraction from gray-level co-occurrence matrix (GLCM) is one of the basic holistic studies that has inspired many face recognition algorithms. This paper presents a theoretically simple, yet efficient, holistic approach that utilizes the spatial relationships of the same pixel patterns occurring at different positions in an image rather than their occurrence statistics as applied in GLCM-based counterparts. The matrix holding the statistical values for the total displacement of the pixel patterns is called the gray-level total displacement matrix (GLTDM). Three approaches are proposed for feature extraction. In the first approach, classical Haralick features extraction is conducted. The second approach (D_GLTDM) utilizes the GLTDM directly as the feature vector rather than extra feature extraction process. In the last approach, principle component analysis (PCA) is used as the feature extraction method. Comprehensive simulations are conducted on images retrieved from the popular face databases, namely face94, ORL, JAFFE and Yale. The performance of the proposed method is compared with that of GLCM, local binary pattern and PCA used in the leading studies. The simulation results and their comparative analysis show that D_GLTDM exhibits promising results and outperforms the other leading methods in terms of classification accuracy.

Haralick texture feature analysis for quantifying radiation response heterogeneity in murine models observed using Raman spectroscopic mapping.

Tumour heterogeneity plays a large role in the response of tumour tissues to radiation therapy. Inherent biological, physical, and even dose deposition heterogeneity all play a role in the resultant observed response. We here implement the use of Haralick textural analysis to quantify the observed glycogen production response, as observed via Raman spectroscopic mapping, of tumours irradiated within a murine model. While an array of over 20 Haralick features have been proposed, we here concentrate on five of the most prominent features: homogeneity, local homogeneity, contrast, entropy, and correlation. We show that these Haralick features can be used to quantify the inherent heterogeneity of the Raman spectroscopic maps of tumour response to radiation. Furthermore, our results indicate that Haralick-calculated textural features show a statistically significant dose dependent variation in response heterogeneity, specifically, in glycogen production in tumours irradiated with clinically relevant doses of ionizing radiation. These results indicate that Haralick textural analysis provides a quantitative methodology for understanding the response of murine tumours to radiation therapy. Future work in this area can, for example, utilize the Haralick textural features for understanding the heterogeneity of radiation response as measured by biopsied patient tumour samples, which remains the standard of patient tumour investigation.

Automatic Counting of Intra-Cellular Ribonucleo-Protein Aggregates in Saccharomyces cerevisiae Using a Textural Approach.

In the context of microbiology, recent studies show the importance of ribonucleo-protein aggregates (RNPs) for the understanding of mechanisms involved in cell responses to specific environmental conditions. The assembly and disassembly of aggregates is a dynamic process, the characterization of the stage of their evolution can be performed by the evaluation of their number. The aim of this study is to propose a method to automatically determine the count of RNPs. We show that the determination of a precise count is an issue by itself and hence, we propose three textural approaches: a classical point of view using Haralick features, a frequency point of view with generalized Fourier descriptors, and a structural point of view with Zernike moment descriptors (ZMD). These parameters are then used as inputs for a supervised classification in order to determine the most relevant. An experiment using a specific Saccharomyces cerevisiae strain presenting a fusion between a protein found in RNPs (PAB1) and the green fluorescent protein was performed to benchmark this approach. The fluorescence was observed with two-photon fluorescence microscopy. Results show that the textural approach, by mixing ZMD with Haralick features, allows for the characterization of the number of RNPs.

Robust and effective clothes recognition system based on fusion of Haralick and HOG features

Robust and effective clothes recognition system based on fusion of Haralick and HOG features

Robust and effective clothes recognition system based on fusion of Haralick and HOG features

In today's modern era, when the computer has become a necessity of an individual, shopping has shifted from shop to online shopping. This kind of clothes classification is used for knowing the name of the cloth that we have seen any movie, serial or anywhere else. In this paper, we present an efficient method to recognise the clothes in natural scenes as well as in the cluttered background. The proposed approach includes three phases: extraction of region of interest (ROI); construction of feature vector; classification. We have validated the proposed approach using our dataset which contains cluttered background images as well as on deep fashion standard dataset. The proposed method successfully resolved the issues of misclassification of clothes in the cluttered background with different illumination conditions. Experimental results show that the proposed technique successfully achieved 88.36% clothes recognition rate.

&lt;i&gt;Indic&lt;/i&gt; script identification from handwritten document images

Script identification plays an important role in document image processing especially for multilingual environment. This paper hires two conventional textural methods for recognition of the scripts of the handwritten documents inscribed in different Indic scripts. The first method extracts well-known Haralick features from spatial grey-level dependence matrix (SGLDM) and the second method computes fractal dimension by using segmentation-based fractal texture analysis (SFTA). Finally, a 104-element feature vector is constructed from each page image by these two methods. The proposed technique is then evaluated on a total dataset comprising 360 handwritten document pages written in 12 Indian official scripts namely Bangla, Devanagari, Gujarati, Gurumukhi, Kannada, Malayalam, Manipuri, Oriya, Tamil, Telugu, Urdu and Roman. Experimentations using multiple classifiers reveal that multilayer perceptron (MLP) shows the highest identification accuracy of 96.94%. The encouraging outcome confirms the efficacy of customary textural features to handwritten Indic script identification.

Study of Haralick’s and GLCM texture analysis on 3D medical images

Purpose of the study: Medical field has highly evolved with advancements in the technologies which prove to be beneficial for radiologists and patients for better diagnosis. The era of medical science provides best healthcare solutions with the help of medical images. Till now, 2D MRIs played a prominent role in early detection of disease but with latest technologies taking over the charge, 3D MRIs are highly effective and great in demand nowadays. With the aid of advanced techniques such as edge detection, segmentation and texture analysis on these images, the disease detection may become much easier.Materials and Methods: Texture of any image is recognized by distribution of gray levels in the neighborhood. The Texture Analysis plays an important role in study of medical images. It identifies the prominent features of an image and highlights the same using different feature extraction technique. In this paper, 3D MRI of human brain is considered and texture analysis based on Haralick's and GLCM texture features is performed. Haralick's feature explains the image intensities of each pixel and their relationship with neighborhood pixels. The entire data set consists of 40 brain tumor patients, out of which a sample has been depicted.Results: The analysis of different features such as Contrast, Correlation, Energy, Homogeneity and Entropy is carried out. Conclusion: Further, the study highlights about the highly useful features for early detection of brain tumor disease.