Breast cancer characterization remains a significant and challenging issue in contemporary medicine. Accurately distinguishing between malignant and benign breast lesions is crucial for effective diagnosis and treatment. The anatomical structure of malignant breast ultrasound images is more chaotic than that of benign images due to disease pathologies. However, texture-based analysis alone often fails to identify the extent of chaoticness in malignant breast ultrasound images due to their vague appearance with normal echo patterns, leading to missed diagnoses and increased mortality rates. To address this issue, we proposed an angular feature-based multilevel breast cancer classification framework mBCCf that aims to improve the accuracy and efficiency of classification. The proposed framework mimics the radiologist interpretation procedure by identifying the chaoticness on the periphery of the breast lesion in a breast ultrasound image (level-1). If the lesion contains an acute angle in any part of the periphery, it can be characterized as malignant or otherwise benign. However, solely relying on level-1 analysis may result in misclassification, especially when benign lesions exhibit echo patterns that resemble malignant ones. To overcome this limitation and to make the proposed system highly sensitive, advanced texture-based analysis (using combined shape, texture, and angular features) is performed (level-2). Finally, the performance of the proposed system is evaluated using a cross-dataset (consisting of 1293 breast ultrasound images) and compared with the different individual feature extraction techniques. Encouragingly, our system demonstrated an accuracy of 96.99% for classifying malignant and benign tumors, which is also validated using statistical analysis. The implications of our research lie in its potential to significantly improve breast cancer diagnosis by providing a reliable, efficient, and sensitive tool for radiologists.

Breast malignancy is a relatively frequent disease that affects people all over the world. When interpreting the lesion component of medical images, inter- and intra-observer errors frequently happen, leading to considerable diversity in result interpretations. To combat this variability, computer-aided diagnosis (CAD) systems are essential. Automatic segmentation is an essential and critical step in CAD systems toward boundary detection, feature extraction, and classification. The aim of this study is to incorporate an Ant colony optimization (ACO) to initialize the cluster center and replace the Euclidean distance (ED) with the Manhattan distance (MD), in the traditional K-means algorithm to segment the BUS images with maximal area preservation. The Jaccard index (JI), Dice similarity (DS), and Area difference (AD) are the cluster validation measures used to compare the efficiency of the proposed method with other state-of-the-art methods. A total of 1293 BUS images are used in this study. According to the quantitative experimental findings, the suggested method can successfully segment the BUS images with an accuracy of 91.66%. Compared to existing methods, the proposed approach accomplishes segmentation more quickly and accurately.

Image captioning is implemented using Deep learning and NLP (Natural Language Processing) resulting in producing a description of an image. The proposed model generates a caption for an image using a Convolutional Neural Network (CNN) together with a Recurrent Neural Network (RNN) and area of attention. Previously, the image names were used as keys to map the images with descriptions. In order to achieve high performance, in the proposed model the image caption is based on the relationship between the areas of a picture (attention model), the words used in the caption, and the state of an RNN language model. The approach of progressive loading is employed for the loading of the image dataset. Further, for encoding the image dataset into a feature vector, VGG16 a pre-trained CNN is used. The extracted feature vector is given as input to the RNN model. These image encodings are output to a specific type of RNN model known as Long Short-Term Memory (LSTM) networks. Subsequently, the LSTM works on decoding the feature vector and predicts the sequence of words, resulting in the generation of descriptions or captions. The training performance is measured using one of the model's quantitative analysis metrics known as BLEU.

Automation has made tremendous changes in the industries. It has been used to automate the manual processes involved in different physical units of the industries. The purpose was to increase the production in the manufacturing. Now, Computers are being used in the industries to monitor functionalities of different production units with the help of artificial intelligence and internet of things (IoT). The IoT has revolutionized the industries. It is an interconnected network system of physical units. The core purpose of it to gather and share information among different physical units. The IoT has great impact on the many areas such as business, industry, medicine, the economy, transport, industrial robots and automation systems. IoT with artificial intelligence has wide range of industrial applications. Industry 4.0 is used in the industries where different industrial units are connected over the internet and interacting to make decisions via machine-to-machine communication. It has increased the benefits of industries in terms of production and supply chain management. Manufacturing industry monitors its production units in every 10 milliseconds to capture features of the product that is being produced. The features generated in this process are huge in amount. Critical observation is performed on the generated features to categorize the product as anomalous or good. Product classification is difficult task in the labeled datasets due to human bias in labeling the final product as anomalous or good. In this work, machine learning models is being used to detect and classify faulty product produced by manufacturing industry. Both qualitative and quantitative study will be carried out to compare various machine learning models.

Feature extraction and classification plays a crucial role in the automated analysis of breast ultrasound (BUS) images. Due to varying sonographic characteristics of benign and malignant lesions, the texture and shape features are mostly employed for designing computer-aided diagnosis (CAD) systems of BUS images. The existing CAD systems use features that are extracted either from the lesion segmented area obtained through segmentation techniques or a rectangular region of interest (ROI) extracted under the guidance of expert Radiologists. However, the significance of features extracted from region comprising only the lesion area is still little explored. This paper investigates the significance of features extracted from the lesion area, lesion surrounding area and rectangular ROI for classification of BUS images. The experiments were conducted on the database of 294 BUS images (104 benign and 190 malignant). Initially, the acquired BUS images were preprocessed through speckle reducing anisotropic diffusion (SRAD) for speckle noise removal. The preprocessed images are segmented using a hybrid segmentation approach including a combination of region-based active contour driven by region-scalable fitting (RBACM-RSF) model and multi-scale Gaussian kernel fuzzy c-means clustering with spatial bias correction (MsGKFCM_S) for getting ROI confined area. The segmented images were further partitioned into two parts (lesion area and lesion surrounding area). Subsequently, a total of 457 texture and shape attributes were extracted from within the lesion area, lesion surrounding area and rectangular ROI comprising of both lesion and its surrounding area. The significance of these features is evaluated using different classifiers (i.e. support vector machine (SVM), Back-propagation artificial neural network (BPANN), Random Forest, AdaBoost). The results indicate that features extracted from within lesion area achieve a maximum classification accuracy of 98.980% with the lowest computational time when linear kernel-based SVM is used.

Reliability and accuracy is the key concern of an automated image classification process. However, the impact of background or surrounding area is very less in compared to object features, which create ambiguity while assigning the appropriate class label and reduce the classification accuracy. This paper presents a new model to address this issue which select the relevant features from the segmented images based on the inner and outer regions. The key idea of this model is that the texture features within the objects are more relevant than the outside area of the objects. The proposed model applying a segmentation method for automated segment the image. The segmented images are then subdivided into two parts (i.e. inner and outer). The 463 shape and texture features are extracted from the inner, outer parts of the segmented images and also from the whole image. Next, these extracted features are used to train the classifier using support vector machine (SVM). A database of 644 images that consisting of 8 classes is used to verify the efficacy of the proposed model. The result proves the efficacy of the proposed model which achieves classification accuracy up to 97.79 % from the inner part of the image. The classification accuracy of inner features is increased by 9.58% from surroundings features.

Ultrasound imaging is most popular technique used for breast cancer screening. Lesion segmentation is challenging step in characterization of breast ultrasound (US) based Computer Aided Diagnosis (CAD) systems due to presence of speckle noise, shadowing effect etc. The aim of this study is to develop an automatic lesion segmentation technique in breast US with high accuracy even in presence of noises, artifacts and multiple lesions. This article presents a novel clustering method called Multi-scale Gaussian Kernel induced Fuzzy C-means (MsGKFCM) for segmentation of lesions in automatically extracted Region of Interest (ROI) in US to delimit the border of the mass. Further, a hybrid approach using MsGKFCM and Multi-scale Vector Field Convolution (MsVFC) is proposed to obtain an accurate lesion margin in breast US images. Initially, the images are filtered using speckle reducing anisotropic diffusion (SRAD) technique. Subsequently, MsGKFCM is applied on filtered images to segment the mass and detect an appropriate cluster center. The detected cluster center is further used by MsVFC to determine the accurate lesion margin. The proposed technique is evaluated on 127 US images using measures such as Jaccard Index, Dice similarity, Shape similarity, Hausdroff difference, Area difference, Accuracy, F-measure and analysis of variance (ANOVA) test. The empirical results suggest that the proposed approach can be used as an expert system to assist medical professionals by providing objective evidences in breast lesion detection. Results obtained are so far looking promising and effective in comparison to state-of-the-art algorithms.

Automatic boundary detection is a challenging and one of the important issues in medical imaging. Contouring breast tumor lesions automatically may avail physicians for correct and faster diseases diagnoses. The ultrasound images are noisy, and boundary detection is a challenging task due to low contrast. The aim of this study is to implement the watershed algorithm in breast cancer ultrasound images to extract precise contours of the tumors. In this process, preprocessing filter reduces the noise by preserving the edges of the tumor lesion. Background and foreground area is calculated based on the threshold. A connected component graph is used to calculate region of interest based on the difference between background and foreground area. Finally, the watershed algorithm is applied to determine the contours of the tumor. In diagnosis applications, automatic lesion segmentation can save the time of a radiologist.

Automated segmentation of tumors in breast ultrasound (US) images is challenging due to poor image quality, presence of speckle noise, shadowing effects and acoustic enhancement. This paper improves the multi-scale Gaussian kernel induced fuzzy C-means clustering method with spatial bias correction (MsGKFCM_S). Furthermore, it presents a hybrid segmentation method, using both the features of the MsGKFCM_S clustering and active contour driven by a region-scalable fitting energy function. The result obtained from the MsGKFCM_S method is utilised to initialise the contour that spreads to identify the estimated regions. It also helps to estimate the several controlling parameters of the curve evolution process. The proposed approach is evaluated on a database of 127 breast US images consisting of 75 malignant and 52 solid benign cases. The performance of proposed approach is compared with other related techniques, using performance measures such as Jaccard Index, dice similarity, shape similarity, Hausdroff difference, area difference, accuracy and F-measure. Results indicate that the proposed approach can successfully detect lesions in breast US images, with high accuracy of 97.889 and 97.513%. Moreover, the proposed approach has the capability of handling shadowing effects, acoustic enhancement and multiple lesions.

With advancements in machine learning algorithms and computer aided diagnostic (CAD) systems, the performance of automated analysis of radiological images has improved substantially in recent times. However, the lack of integration between the radiologist and CAD systems restrains the rate of progress as well as the reach of such advancements in clinical use. This article aims to improve the clinical efficiency of ultrasound based CAD systems for classification of breast lesions by integrating back-propagation artificial neural network (BPANN), support vector machine (SVM) and radiologist feedback. The acquired breast ultrasound images were subjected to wavelet based filtering in order to reduce speckle noise followed by feature extraction, feature selection and classification. Experiments on a database of 178 ultrasound images of breast anomalies (88 benign and 90 malignant) show that the proposed methodology achieves classification accuracy of 98.621% and 98.276%, respectively, when all 457 and 19 most relevant features selected by multi-criteria feature selection method were used for classification. The accuracy achieved is significantly higher than that using conventional classifiers based on BPANN and SVM. Further, it is found that integrating expert opinion in CAD systems improves its overall performance. The quantitative results obtained are discussed in light of some recently reported studies.

Automatic seed selection is an important and crucial step toward the boundary detection in ultrasound B-scan images. This paper focuses on a methodological framework that can automatically detect a seed point of an ultrasound image by using texture features. Based on the selected seeds of cluster the ultrasound images are segmented using active contour, K-means and Otsu methods. The comparative analysis of these segmentation techniques is also reported. The proposed method is applied on 116 ultrasound images in which 45 are benign cases and 71 malignant cases. The quantitative experimental results show that the proposed method can successfully find an accurate seed point based on texture features and it has the ability to segment the image with high accuracy of 89.65 %. The proposed method is faster and performs more accurate segmentation than existing algorithms.

Missing values can cause serious problems while mining data sets, such as i) loss of information and efficiency; ii) problem in data handling computation and analysis due to irregularities in the data patterns and non-applicability of standard software; and iii) serious bias if there are systematic differences between the observed and the unobserved data. Missing values can also cause misleading results by introducing bias. This paper focuses on a methodological framework for the development of an automated data imputation model based on Hybrid Higher Order Neural Network Classifier (HHONC). Four real, integer and simulated data sets are exposed to a perturbation experiment, based on the random generation of missing values. Here different imputation methods are applied in glass identification, wine recognition, heart disease and lung cancer data set to find the missing value and compared with different classic imputation procedures. The experiment not only improves the quality of a database with missing value but also the best results are clearly obtained with different variables.

Data sets contain very large amount of data which is not an easy task for the user to scan the entire data set. The researcher's initial task is to formulate a rational justification for the use of sampling in his research. Sampling has been often suggested as an effective tool to reduce the size of the dataset operated at some cost to accuracy. It is the process of selecting representatives which indicates the complete data set by examining a fraction. Due to sampling we overcome the problems like; i) in research it is not possible to collect and test each and every element from the data base individually; and ii) study of sample rather than the entire dataset is also sometimes likely to produce more reliable results. This paper focuses on different types of sampling strategies applied on neural network. Here sampling technique has been applied on two real, integers and categorical dataset such as yeast and hepatitis data set prior to classification. The main objective of this paper is an empirical comparison of different sampling strategies for classification which gives more accuracy. © 2012 Published by Elsevier Ltd.

Data sets contain very large amount of information, which is not an easy task for the users to scan the entire data set. The researcher's initial task is to formulate a realistic explanation for the use of sampling in his research. Sampling has been often suggested as an effective tool to reduce the size of the dataset operated at some cost to accuracy. It is the the process of selecting a representative part of a data set for the purpose of determining parameters or characteristics of the whole data set. Due to sampling we overcome the problems like; i) in research it is not possible to collect and test each and every element from the data base individually; and ii) study of sample rather than the entire dataset is also sometimes likely to produce more reliable results. This paper focuses on different types of sampling strategies applied on hybrid higher order neural network classifier (HHONC) rather than artificial neural network which is having several limitations. To overcome such limitations HHONC have been used. Here sampling technique has been applied on four real, integers and categorical dataset such as breast cancer, pima Indian diabetes, leukaemia and lung cancer data set prior to classification. The main objective of this paper is an empirical comparison of different sampling strategies for classification which gives more accuracy. © 2012 ACM.

Missing data are common occurrences and can have a significant effect on the conclusions that can be drawn from the data. In statistics, missing data or missing values occur when no data value is stored for the variable in the current observation. Due to missing value we are facing several problems like information loss for computation and analysis of data. Missing values can also cause misleading results by introducing bias. Serious bias is a systematic difference between the observed and the unobserved data. This paper focuses on a methodological framework for the development of an automated data imputation model based on wavelet neural network (WNN). Here we use an adaptive higher order functions or different wavelet functions as the kernel of NN instead of each neuron activation function. A wavelet is a wavelike oscillation with a amplitude that starts out at zero, increases, and then decreases back to zero. Generally, wavelets are purposefully crafted to have specific properties that make them useful for signal processing. Six real, integer and simulated data sets are exposed to a perturbation experiment, based on the random generation of missing values. Here neural network (NN) and WNN is applied in glass identification, wine recognition, heart disease, leukemia, breast cancer and lung cancer data set to find the missing value and compared with different classic imputation procedures. The experiment conducted considering different performance measures using WNN, not only improves the quality of a database with missing value but also the best results are clearly obtained with different variables. © 2012 ACM.