Current Signal Transduction Therapy - Volume 11, Issue 2, 2016

Background: Digital Mammography is the most vital and successful imaging modality used by radio diagnosis method to find out breast cancer. Breast cancer is the most significant and common cause of cancer death in women. The main problem is to find the accurate and efficient method for breast cancer segmentation. Method: The morphological method is the most important approach in image segmentation method. There are various new methods available for breast cancer image segmentation but those methods are not upto the mark. They fall behind the image segmentation. Results: On comparing both the algorithms for segmenting the mammographic images, applying the Neural Networks algorithm will be a better option rather than applying Region Growing Algorithm. The accuracy of the segmentation is higher in the morphological image segmentation approach. Conclusion: The results show that, the performance of morphological approach is more efficient than other methods.

Background: Colon cancer remains among the top perpetrators of deaths linked to cancer. The probability of cancer reaching more parts of the body is extremely high in colorectal cancers. Early detection is hence, highly important for faster treatment. Method: In the current work, a hybrid approach toward the detection of colon cancers through the usage of microarray datasets, is presented. Particle Swarm Optimization (PSO) is utilized for features extraction, while Support Vector Machine (SVM) and Bagging approaches are utilized as classifiers. Results: The Colon Microarray Gene Dataset is used to evaluate minimum Redundancy Maximum Relevance (mRMR), Bagging, SVM, PSO and PSO-SVM with regard to classification accuracy, sensitivity and specificity. The proposed PSO-SVM displays best performance in all categories. Conclusion: Experiments reveal the capabilities of the proposed PSO-SVM to explore features space for the optimal features combination for gene selection from microarray data.

Background: Microarray technologies provide huge amount of information and is particularly helpful in the prediction and diagnosis of cancer. To accurately classify cancers, genes related to cancer have to be selected, as genes mined from microarrays possess too much noise. Method: In the current work, new binary modifications of the Grey Wolf Optimization (GWO) is suggested for choosing optimal features subsets for classification. In the proposed approach, GWO is modified by binarizing only the initial three optimal solutions and updation of the wolf position using stochastic crossover. Modification was also carried out using sigmoidal functions to compress the continuous updated positions. Multilayer Perceptron – Neural Network (MLP-NN) classifier is used for classifying the selected features. Results: The ALL/AML Leukemia dataset is used for evaluating Markov Blanket filter, minimum Redundancy Maximum Relevance (mRMR), Binary GWO and Mutated Binary GWO (MBGWO) with regard to classification accuracy, sensitivity and specificity. The proposed MBGWO achieved classification accuracy of 95.45% and also has better sensitivity and specificity. Conclusion: Experiments reveal the capabilities of the proposed MBGWO to explore features space for the optimal features combination for gene selection from microarray data.

Background: Mammographies are a significant technology utilized in the effective detection of breast cancers prior to them becoming palpable during selfexaminations. The primary aim of the study was the determination of screening precision of mammographies as well as ultrasounds in local populations. Method: In the current study, Minimum Spanning Tree (MST) segmentations are suggested for the selection of minute weights amongst all spanning trees. In the suggested method, Fuzzy Local Binary Patterns (FLBPs) are images comprising micro-patterns. LBPs are first-order circular derivatives of patterns which are created through the concatenation of binary gradient directions. It includes fuzzy logic in LBPs through sets of fuzzy rules. Support Vector Machines (SVMs) are utilized for the classification of chosen attributes. Results: Breast cancers in ultrasounds are utilized for the valuation of KNN, SVM-0.1, Naïve Bayes, SVM-0.5 as well as SVM-0.3 approaches in terms of classification precision, sensitivities, specificities, Positive Predictive Values (PPVs) as well as Negative Predictive Values (NPVs). The suggested that SVM-0.3 showed the most optimal performance in all factors. Conclusion: Breast imaging utilizing mammographies as well as sonographies among women who display local or diffused breast pains are of considerable importance, for the assurance of patients as well as clinicians. But if imaging discoveries are symptomatic of pathologies, biopsies ought not to be put off for long.

Background: Neural Networks are utilized in several applications in the field of healthcare, one such being the classification of lung cancers. Innovative advancements in diagnosing tumours are a major boost to developing novel treatment techniques in the early stages of lung cancer. Method: In this work, a novel image-based features selection method for classifying lung Computed Tomography (CT) images is introduced. A new fusion-based technique through a combination of Gabor filters and first order histograms was created. The suggested model utilizes Multi-Layer Perceptron Neural Networks (MLP-NN) alongside Krill Herds (KH) for structural optimization which consists of three phases. First, images are pre-processed and features selected through the new fusion-based selection method. Next, the selected features are got through the application of Correlation based Feature Selection (CFS), Mutual Information (MI), Fuzzy Unordered Rule Induction Algorithm (FURIA) that choose the highest ranked ones. Lastly, classifiers like AdaBboost or MLP-NN carry out the classification of the cancers. Results: Misclassification rates, average true positive rates, average false discovery rates are utilized in the evaluation of CFS-Furia, CFS-AdaBoost classifier, CFS-MLPNN, CFS-KHMLPNN, MI-Furia, MIAdaBoost classifier, MI-MLPNN and MI-KHMLPNN. The suggested MI-KHMLPNN outperforms all others in every category. Conclusion: The model was evaluated with several lung CT images and has proven to attain excellent results in the classification of lung cancers.

Background: Breast cancers are one of the most prevalent forms of cancers amongst women apart from being the top second cause of cancer related deaths across the world. A way to detect the presence of breast cancers earlier is through the presence of minute deposits of calcium, that is, micro-calcifications in mammograms. Detecting this much earlier is important for successfully treating the cancer. Method: In the current work, a novel Hybrid Particle Swarm Optimized (PSO) cascaded classifier is suggested. Wavelet Features linked to Histogram of Oriented Gradients (HOG) based features selection technique is utilized. For the detection of micro-calcifications utilizing the Support Vector Machines (SVM) is a supervised learning technique which may be utilized for classifications as well as regressions. Cascade classifications are capable of tackling all the mentioned challenges with an integrated model through the usage of asymmetric cascades of sparse classifiers, every single one trained to attain great detection sensitivities as well as adequate false positive rates. Results: SVM-Poly kernels, SVM-Radial Basis Function (RBF) kernels, Cascaded classifiers, PSO optimized cascade classifiers as well as Hybrid PSO optimized cascade classifiers are valuated for classification accuracies, sensitivities as well as specificities. The suggested Hybrid PSO optimized cascade classifiers outperform all other in all categories. It was observed that the proposed Hybrid PSO optimized cascade classifier method increased classification accuracy by 2.49%-8.74%, compared with other methods. Conclusion: Experimental evaluations proved the abilities of the suggested Hybrid PSO optimized cascade classifiers for the exploration of features space for best features combination for malignant and benign micro-calcifications from mammograms.

Background: Breast cancers are conventionally the leading cause of cancer-related deaths amongst women. For the reduction of death rates by earlier identification of carcinogenic areas, mammogram images are utilized. Computer aided diagnosis plays an important role in screening of the mammograms. Mammography is an efficient as well as feasible method for the detection of breast cancers, especially minute tumours. Efficient performance of these tools is dependent on the efficacy of the classifier algorithms. In this work, feature selection techniques are proposed to improve the efficacy of the classifiers. Method: The major phases in diagnosing breast cancers are features extraction and selection. Detecting tumours naturally requires extraction of features as well as their classification. This work uses Pseudo Zernike Moments and Gaussian Markov Random Field (GMRF) for feature extraction, Binary Shuffled frog algorithm, Information Gain (IG) and Binary Particle Swarm Optimization (PSO) for feature selection and C4.5, Random tree, Adaboost as classifiers. Results: The use of feature selection techniques for successfully selecting relevant feature subset improves the classification performance. The proposed wrapper based Binary Shuffled Frog Algorithm improves the detection of breast cancer mass in mammograms. Conclusion: The work focuses on improving classification performance through feature selection. The experimental results demonstrate the efficacy of the proposed feature selection method. It is observed that the feature selectors are necessary to improve the efficiency of the classifiers. It is observed that among the various classification techniques, C4.5 outperforms other algorithms achieving the highest accuracy.

Applications of medical images are increasing rapidly, moreover the medical images produced by the imaging devices occupies more volume of space. Medical information system needs to store huge amount of data in the form of medical images for future record and also these images are required to be transmitted over network to the place of specialist to obtain diagnostics opinion. Because of the limited availability of space and bandwidth, an efficient compression technique is needed to reduce the bits to store and transmit these images. In this paper, we propose an efficient scheme of medical image compression based on region of interest using Peano space filling curve. In this method, region containing the most useful diagnostic features are treated as region of interest (ROI), pixels in ROI region are arranged using Peano space filling curve (PSFC) and entropy encoded without any loss in quality. The remaining regions are treated as non region of interest(NROI) and are encoded with singular value decomposition followed by entropy encoding. The encoded ROI and NROI region are combined to give the compressed output. The performance of the proposed compression technique with various encoding such as Huffman and LZW along with Peano Space Filling curve reordering is compared with standard JPEG2000 compression. The result shows that proposed method gives better performance in terms of PSNR and Compression Ratio.

Background: Automated Medical Image Analysis has emerged as an important tool for the diagnoses of anatomical pathology and can be integrated with the medical information system to deliver useful information for the health care provider. Method: This study proposes a novel SVM Classifier Using Soft Computing Approach for Automated Classification of Emphysema, Bronchiectasis and Pleural Effusion Using Optimized Gabor Filter. In this work an improved Gabor Filter using Firefly optimization algorithm to find the optimal Gabor parameters is proposed. The proposed technique is used to extract features from lung CT images and classify automatically the images as Normal, Emphysema, Bronchiectasis and Pleural Effusion using SVM classifier. Results: 40 CT images of Emphysema, Bronchiectasis, pleural effusion and 100 CT images of normal are used. The performance of classification accuracy, average PPV, average sensitivity, and average f measure is used for evaluating all the techniques. The proposed Optimized Gabor Filter -SVM displays best performance in all categories. Conclusion: The proposed method was tested with a number of CT lung images and satisfactory results was achieved in classifying the lung diseases.

Background: Segmenting images is the most difficult as well as a happening research topic in the domain of image processing. Despite accessibility to a huge number of excellent approaches for brain Magnetic Resonance Imaging (MRI) segmentations, it remains a difficult job and speed of the technique requires great improvements. Medical images segmentations need effective as well as strong segmentation models that are robust against noisy data. Method: In the current work, the suggested Fuzzy C-Means (FCM) is a widely utilized approach to segment medical images though it regards merely image intensity and hence, provides poor outcomes for noise-filled images. Classifiers like Bagging as well as Boosting are common resampling ensemble approaches creating and merging several classifiers through the same learning model for base classifiers. Boosting models are more robust than bagging ones on noiseless data. Boosting decreases errors of weak learning models which create classifiers that are merely a little better than arbitrary guesses. Results: Results show that the Fuzzy Bee Segmentation Bagging method increased classification accuracy by 4.34%, 3.02% & 1.71% when compared with FCM Segmentation- Boosting, FCM Segmentation - Bagging and Fuzzy Bee Segmentation - Boosting methods. Conclusion: The final images with their details are very much helpful for further brain MRI image processing and analysis in medical diagnosis. Results show that the classification accuracy, precision as well as recall, was better than all other methods regarded for experiments.