A Multiscale Hierarchical Threshold-Based Completed Local Entropy Binary Pattern for Texture Classification Abstract Over the year, visual texture analysis has come to be recognized as one of the most important methods in the area of medical image analysis and understanding, face description and detection, and so on. The goal of texture descriptors is to capture the general characteristic of textures such as dependency as well as invariance properties. Among all the texture descriptors, the binary pattern family of algorithms achieves a great trade of representation efficiency and complexity. This work introduces an efficient discriminative texture descriptor for visual texture classification. Its main contribution is twofold: a multiscale thresholding framework based on hierarchical adaptive local partition to binary encoding and an efficient completed local entropy binary pattern (CLEBP) descriptor. The basic completed local entropy binary pattern is extended by multiscale thresholding framework with hierarchical thresholding to capture not only microstructure local patterns but also macrostructure texture information. Such extension improves the quality and discriminative factor of texture classification. Extensive experiments on three widely used benchmark texture databases (Outex, UIUC, and KTH-TIPS) proof the efficiency of the proposed visual texture descriptor and hierarchical thresholding strategy. Compared with some classical local binary pattern variants and many state-of-the-art methods, the proposed descriptor achieves competitive and superior texture classification performance. The results prove that the proposed method is a powerful and effective texture descriptor for visual texture classification.

A Multicriteria Decision-Making Approach with Linguistic D Numbers Based on the Choquet Integral Abstract Linguistic D numbers (LDNs) provide a reliable expression of cognitive information. By inheriting the advantages of linguistic terms (LTs) and D numbers (DNs), LDNs can express uncertain and incomplete cognitive information in multicriteria decision-making (MCDM), and they do so better than existing methods. The TODIM (an acronym in Portuguese of interactive and multicriteria decision-making) method can consider decision experts’ (DEs’) bounded rationality, such as cognition toward loss, which is caused by the DEs’ cognitive limitations during the decision process. Additionally, the Choquet integral can process the interrelationship among criteria or cognitive preferences, which helps to reflect the complex cognition of DEs. Therefore, it is necessary to propose a novel cognitive MCDM approach by extending the TODIM method and Choquet integral to handle MCDM problems in which the cognitive information is expressed by LDNs. In this paper, we introduced LDNs to represent uncertain and hesitant cognitive information. The definition and comparison approach of LDNs were also recommended. Then, we proposed the distance function and modified the score function of LDNs. Later, considering the limitations of the DEs’ cognitive abilities in real decision-making and the phenomenon where attributes or cognitive preferences in MCDM problems are not independent, we developed a novel cognitive MCDM approach with LDNs by extending the TODIM method and the Choquet integral to deal with these cases. The proposed approach can not only take the influence of the limited cognitive abilities of DEs on the decision-making results into account but can also deal with the correlation between the cognitive preferences. A novel cognitive MCDM approach with LDNs based on the TODIM method and Choquet integral was proposed. Moreover, the validity and superiority of the presented approach were verified by dealing with practical problems and comparing them to other approaches. The proposed approach can consider cases where the DEs are rationally bounded in their cognitive decision-making and the criteria or cognitive preferences in MCDM problems have an interrelationship. Therefore, this approach can produce more reliable decision-making results than some existing MCDM approaches.

Improvements on Correlation Coefficients of Hesitant Fuzzy Sets and Their Applications Abstract Hesitant fuzzy set (HFS) can express the hesitancy and uncertainty according to human’s cognitions and knowledge. The decision making with HFSs can be regarded as a cognitive computation process. Decision making based on information measures is a hot topic, among which correlation coefficient is an important direction. Although many correlation coefficients of HFSs have been proposed in the previous papers, they suffer from different counter-intuitions to a certain extent. Therefore, we mainly focus on improving these counter-intuitions of the existing correlation coefficients of HFSs in this paper. We point out the counter-intuitions of the existing correlation coefficients of HFSs and analyze the reasons of them in the view of the rigorous mathematics and stochastic process rules. We improve these counter-intuitions and develop the correct versions. Moreover, we use two examples about medical diagnosis and cluster analysis to compare the improved correlation coefficients with the existing ones. The improved correlation coefficients can handle the examples well. Further, combining with the comparison analysis, the accuracy and discrimination property of the improved correlation coefficients are demonstrated in detail, which shows the advantages of them. The notion of the improved correlation coefficients can benefit other types of fuzzy sets too.

Cognitively Inspired Feature Extraction and Speech Recognition for Automated Hearing Loss Testing Abstract Hearing loss, a partial or total inability to hear, is one of the most commonly reported disabilities. A hearing test can be carried out by an audiologist to assess a patient’s auditory system. However, the procedure requires an appointment, which can result in delays and practitioner fees. In addition, there are often challenges associated with the unavailability of equipment and qualified practitioners, particularly in remote areas. This paper presents a novel idea that automatically identifies any hearing impairment based on a cognitively inspired feature extraction and speech recognition approach. The proposed system uses an adaptive filter bank with weighted Mel-frequency cepstral coefficients for feature extraction. The adaptive filter bank implementation is inspired by the principle of spectrum sensing in cognitive radio that is aware of its environment and adapts to statistical variations in the input stimuli by learning from the environment. Comparative performance evaluation demonstrates the potential of our automated hearing test method to achieve comparable results to the clinical ground truth, established by the expert audiologist’s tests. The overall absolute error of the proposed model when compared with the expert audiologist test is less than 4.9 dB and 4.4 dB for the pure tone and speech audiometry tests, respectively. The overall accuracy achieved is 96.67% with a hidden Markov model (HMM). The proposed method potentially offers a second opinion to audiologists, and serves as a cost-effective pre-screening test to predict hearing loss at an early stage. In future work, authors intend to explore the application of advanced deep learning and optimization approaches to further enhance the performance of the automated testing prototype considering imperfect datasets with real-world background noise.

Multi-Granulation-Based Graphical Analytics of Three-Way Bipolar Neutrosophic Contexts Abstract Recently, a three-way fuzzy concept lattice and its graphical structure analytics has given a mathematical way to deal with cognitive concept learning based on its truth, false, and uncertain regions, independently. In this process, a major problem was addressed while existence of bipolar information in a three-way decision space. To address this problem, the current paper aimed at introducing bipolar neutrosophic graph representation of concept lattice and its granular-based processing for cognitive concept learning. In addition, the proposed method is illustrated with an example for better understanding. Cognitive computing provides a way to mimic with human brain and its uncertainty beyond the binary values. To characterize these types of bipolar attributes based on its acceptation, rejection, and uncertain part, the three-way bipolar neutrosophic context and its concept lattice is introduced in this paper. In addition, another method is proposed to extract some of the bipolar cognitive concepts based on user required bipolar truth, bipolar indeterminacy, and falsity membership values, independently. This paper provides a graphical structure visualization of the three-way bipolar information at user defined granules. It is also shown that the extracted information from both of the proposed methods are concordant with each other. It is also shown that, the proposed method provides an adequate way to model the three-way bipolar cognitive concepts when compared to other available approaches. This paper introduces a method to model the three-way bipolar cognitive context using the properties of bipolar neutrosophic graph and its lattice structure. The line diagram is drawn based on their lower neighbors within O(|C| n2 m3) time complexity. In addition, another method is proposed to refine the three-way bipolar neutrosophic cognitive concepts at user defined granulation within O(n6) or O(m6) time complexity with an illustrative example. However, the proposed method is unable to measure the changes in the three-way bipolar neutrosophic cognitive concepts at the given phase of time. Due to that, the author will focus on resolving this issue of bipolar neutrosophic context in near future.

A Handwriting-Based Protocol for Assessing Neurodegenerative Dementia Abstract Handwriting dynamics is relevant to discriminate people affected by neurodegenerative dementia from healthy subjects. This can be possible by administering simple and easy-to-perform handwriting/drawing tasks on digitizing tablets provided with electronic pens. Encouraging results have been recently obtained; however, the research community still lacks an acquisition protocol aimed at (i) collecting different traits useful for research purposes and (ii) supporting neurologists in their daily activities. This work proposes a handwriting-based protocol that integrates handwriting/drawing tasks and a digitized version of standard cognitive and functional tests already accepted, tested, and used by the neurological community. The protocol takes the form of a modular framework which facilitates the modification, deletion, and incorporation of new tasks in accordance with specific requirements. A preliminary evaluation of the protocol has been carried out to assess its usability. Successively, the protocol has been administered to more than 100 elderly MCI and match controlled subjects. The proposed protocol intends to provide a “cognitive model” for evaluating the relationship between cognitive functions and handwriting processes in healthy subjects as well as in cognitively impaired patients. The long-term goal of this research is the development of an easy-to-use and non-invasive methodology for detecting and monitoring neurodegenerative dementia during screening and follow-up.

Improving User Attribute Classification with Text and Social Network Attention Abstract User attribute classification is an important research topic in social media user profiling, which has great commercial value in modern advertisement systems. Existing research on user profiling has mostly focused on manually handcrafted features for different attribute classification tasks. However, these research has partly overlooked the social relation of users. We propose an end-to-end neural network model called the social convolution attention neural network. Our model leverages the convolution attention mechanism to automatically extract user features with respect to different attributes from social texts. The proposed model can capture the social relation of users by combining semantic context and social network information, and improve the performance of attribute classification. We evaluate our model in the gender, age, and geography classification tasks based on the dataset from SMP CUP 2016 competition, respectively. The experimental results demonstrate that the proposed model is effective in automatic user attribute classification with a particular focus on fine-grained user information. We propose an effective model based on the convolution attention mechanism and social relation information for user attribute classification. The model can significantly improve the accuracy in various user attribute classification tasks.

Facial Expression Recognition Based on a Hybrid Model Combining Deep and Shallow Features Abstract Facial expression recognition plays an important role in the field involving human-computer interactions. Given the wide use of convolutional neural networks or other neural network models in automatic image classification systems, high-level features can be automatically learned by hierarchical neural networks. However, the training of CNNs requires large amounts of training data to permit adequate generalization. The traditional scale-invariant feature transform (SIFT) does not need large learning samples to obtain features. In this paper, we proposed a feature extraction method for use in the facial expressions recognition from a single image frame. The hybrid features use a combination of SIFT and deep learning features of different levels extracted from a CNN model. The combined features are adopted to classify expressions using support vector machines. The performance of proposed method is tested using the publicly available extended Cohn-Kanade (CK+) database. To evaluate the generalization ability of our method, several experiments are designed and carried out in a cross-database environment. Compared with the 76.57% accuracy obtained using SIFT-bag of features (BoF) features and the 92.87% accuracy obtained using CNN features, we achieve a FER accuracy of 94.82% using the proposed hybrid SIFT-CNN features. The results of additional cross-database experiments also demonstrate the considerable potential of combining shallow features with deep learning features, and these results are more promising than state-of-the-art models. Combining shallow and deep learning features is effective when the training data are not sufficient to obtain a deep model with considerable generalization ability.

Multi-Region Risk-Sensitive Cognitive Ensembler for Accurate Detection of Attention-Deficit/Hyperactivity Disorder Abstract In this paper, we present a multi-region ensemble classifier approach (MRECA) using a cognitive ensemble of classifiers for accurate identification of attention-deficit/hyperactivity disorder (ADHD) subjects. This approach is developed using the features extracted from the structural MRIs of three different developing brain regions, viz., the amygdala, caudate, and hippocampus. For this study, the structural magnetic resonance imaging (sMRI) data provided by the ADHD-200 consortium has been used to identify the following three classes of ADHD, viz., ADHD-combined, ADHD-inattentive, and the TDC (typically developing control). From the sMRIs of the amygdala, caudate, and hippocampus regions of the brain from the ADHD-200 data, multiple feature sets were obtained using a feature-selecting genetic algorithm (FSGA), in a wraparound approach using an extreme learning machine (ELM) basic classifier. An improved crossover operator for the FSGA has been developed for obtaining higher accuracies compared with other existing crossover operators. From the multiple feature sets and the corresponding ELM classifiers, a classifier-selecting genetic algorithm (CSGA) has been developed to identify the top performing feature sets and their ELM classifiers. These classifiers are then combined using a risk-sensitive hinge loss function to form a risk-sensitive cognitive ensemble classifier resulting in a simultaneous multiclass classification of ADHD with higher accuracies. Performance evaluation of the multi-region ensemble classifier is presented under the following three scenarios, viz., region-based individual (best) classifier, region-based ensemble classifier, and finally a multiple-region-based ensemble classifier. The study results clearly indicate that the proposed “multi-region ensemble classification approach” (MRECA) achieves a much higher classification accuracy of ADHD data (normally a difficult problem because of the variations in the data) compared with other existing methods.

Discriminant Zero-Shot Learning with Center Loss Abstract Current work on zero-shot learning (ZSL) generally does not focus on the discriminative ability of the models, which is important for differentiating between classes since our brain focuses on the discriminating part of the object to classify it. For generalized ZSL (GZSL), the fact that the outputs of the model are not comparable leads to a degraded performance. We propose a new ZSL method with a center loss to make the instances from the same class more compact by extracting their discriminative parts. Further, we introduce a varying learning rate to accelerate the model selection process. We also demonstrate how to boost the performance of GZSL by rectifying the outputs of the model to make the outputs be comparable. Experimental results on four benchmarks, including SUN, CUB, AWA2, and aPY, demonstrate the superiority of the proposed method, therein achieving state-of-the-art performance.