The progress of communications over internet media such as social media and messengers has led to the production of large amount of textual data. This kind of information contains a lot of valuable knowledge and can be used to improve the performance of other natural language processing (NLP) tasks. There are several ways to use such information, of which one is text summarization. Summarizing textual information can extract the main content of text within a short time. In this paper, we propose an approach for extractive summarization on Persian texts by using sentences embedding and a Sparse Coding framework. Most previous works focuses on text’ s sentences individually which may not consider the hidden structure patterns between them. In this paper, our proposed approach can consider the relations between the text’ s sentences by using three main criteria, namely coverage, diversity and sparsity, when selecting the summary sentences. By considering these criteria, we select sentences that can reconstruct the whole text with least reconstruction error. The proposed approach is evaluated on Persian dataset Pasokh and achieved 10. 02% and 8. 65% improvement compared to the state-of-theart methods in rouge-1 and rouge-2 f-scores, respectively. We show that considering semantic relations among the text’ s sentences can lead us to better sentence summarization.

The progress of communications over internet media such as social media and messengers has led to the production of large amount of textual data. This kind of information contains a lot of valuable knowledge and can be used to improve the performance of other natural language processing (NLP) tasks. There are several ways to use such information, of which one is text summarization. Summarizing textual information can extract the main content of text within a short time. In this paper, we propose an approach for extractive summarization on Persian texts by using sentences embedding and a Sparse Coding framework. Most previous works focuses on text’s sentences individually which may not consider the hidden structure patterns between them. In this paper, our proposed approach can consider the relations between the text’s sentences by using three main criteria, namely coverage, diversity and sparsity, when selecting the summary sentences. By considering these criteria, we select sentences that can reconstruct the whole text with least reconstruction error. The proposed approach is evaluated on Persian dataset Pasokh and achieved 10. 02% and 8. 65% improvement compared to the state-of-the-art methods in rouge-1 and rouge-2 f-scores, respectively. We show that considering semantic relations among the text’s sentences can lead us to better sentence summarization.

Due to the development of social networks and the Internet of things, we recently have faced with large datasets. High-dimensional data is mixed with redundant and irrelevant features, so the performance of machine learning methods is reduced. Feature selection is a common way to tackle this issue with aiming of choosing a small subset of relevant and non-redundant features. Most of the existing feature selection works are for supervised applications, which assume that the information of class labels is available. While in many real-world applications, it is not possible to provide complete knowledge of class labels. To overcome this shortcoming, an unsupervised feature selection method is proposed in this paper. The proposed method uses the matrix factorization-based regularized self-Representation model to weight features based on their importance. Here, we initialize the weights of features based on the correlation among features. Several experiments are performed to evaluate the effectiveness of the proposed method. Then the results are compared with several baselines and state-of-the-art methods, which show the superiority of the proposed method in most cases.

Texture and color are two important attributes for object recognition. Recently, quaternionic Representation of color images have been used as an effective method for color image processing. Using such a Representation, it is possible to consider the mutual interaction between different color channels. In the last decade, several quaternion operations like rotation, reflection, and Clifford translation have been developed. Such operators are able to extract shallow information from the color images. In this paper, we first propose a set of new quaternion operators called hybrid quaternionic operators, which can be produced by a cascade of several simple quaternionic operators. Such operators can extract deeper information from the color images. We then use such operators, and present a novel color texture classification method using the concept of Sparse Coding. Experimental results indicate that the proposed method outperforms several existing and popular methods.

Obsessive-Compulsive Disorder (OCD) is the fourth most common mental disorder and the tenth cause of disability worldwide. This disorder can lead to cognitive impariments in attention, memory, thinking, auditory processing of words and visual cognition. Previous studies have demonstrated that OCD is associated with changes in connectivity between different lobes of the brain. Hence, the quantification of symmetry and connectivity between different brain regions has attracted great attention. This study has provided a new efficient approach based on analytic Representation of EEG signals and statistical features to quantify the difference of intrinsic components of brain activity between brain lobes. For this purpose, phase spectra and amplitude envelopes of the analytic EEG signals have been extracted and analyzed. Furthermore, Non-Negative Least Square Sparse classification method has been used for discriminating between healthy control group and OCD patients. The detection capability of the proposed method has been studied in 19 healthy subjects and 11 patients, performing simple flanker tasks. The obtained results have demonstrated the effectiveness of the combined amplitude and phase information in OCD detection with high average accuracy rate of 93.78 %. In comparison between different regions, the inter-hemispheric features and those extracted from the frontal lobe and frontal-parietal network have shown more efficiency in diagnosing the OCD. This study has also highlighted more importance of amplitude information in the OCD detection.

Personal identification based on vein pattern is one of the latest biometric approaches that have attracted lots of attention. Besides, Convolutional Sparse Coding (CSC) is a popular model in the signal and image processing communities, resolving some limitations of the traditional patch-based Sparse Representations. As most existing CSC algorithms are suited for image restoration, we present a novel discriminative model based on CSC for dorsal hand vein recognition in this paper. The proposed method, discriminative local block coordinate descent (D-LoBCoD), is based on extending the LoBCoD algorithm by incorporating the classification error into the objective function that considers the performance of a linear classifier and the Representational power of the filters simultaneously. Thus, for training, in each iteration, after updating the Sparse coefficients and convolutional filters, we minimize the classification error by updating the classifier’s parameters according to the label information. Finally, after training, the label of the query image will be determined by the trained classifier. One thousand two hundred dorsal hand vein images taken from 100 individuals are used to verify the validity of the proposed methods. The experimental results show that our method outperforms other competing methods. Further, we demonstrate that our proposed method is less dependent on the number of training samples because of capturing more representative information from the corresponding images.

Due to the growing increase of generated images via cameras and various instruments, image processing has found an important role in most of practical usages including medical, security and driving. However, most of the available models has no considerable performance and in some usages the amount of error is very effective. The main cause of this failure in most of available models is the distribution mismatch across the source and target domains. In fact, the made model has no generalization to test data with different properties and distribution compared to the source data, and its performance degrades dramatically to face with new data. In this paper, we propose a novel approach entitled Sparse Coding and ADAptive classification (SADA) which is robust against data drift across domains. The proposed model reduces the distribution difference across domains via generating a common subspace between the source and target domains and increases the performance of model. Also, SADA reduces the distribution mismatch across domains via the selection of the source samples which are related to target samples. Moreover, SADA adapts the model parameters to build an adaptive model to encounter with data drift. Our variety of experiments demonstrate that the proposed approach outperforms all stat-of-the-art domain adaptation methods.

Analyzing motion patterns in traffic videos can be exploited directly to generate highlevel descriptions of the video contents. Such descriptions may further be employed in different traffic applications such as traffic phase detection and abnormal event detection. One of the most recent and successful unsupervised methods for complex traffic scene analysis is based on topic models. In this paper, a two-level Sparse Topical Coding (STC) topic model is proposed to analyze traffic surveillance video sequences which contain hierarchical patterns with complicated motions and co-occurrences. The first level STC model is applied to automatically cluster optical flow features into motion patterns. Then, the second level STC model is used to cluster motion patterns into traffic phases. Experiments on a real world traffic dataset demonstrate the effectiveness of the proposed method against conventional onelevel topic model based methods. The results show that our two-level STC can successfully discover not only the lower level activities but also the higher level traffic phases, which makes a more appropriate interpretation of traffic scenes. Furthermore, based on the two-level structure, either activity anomalies or traffic phase anomalies can be detected, which cannot be achieved by the one-level structure.

Image Representation is a crucial problem in image processing where there exist many low-level Representations of image, i. e., SIFT, HOG and so on. But there is a missing link across low-level and high-level semantic Representations. In fact, traditional machine learning approaches, e. g., non-negative matrix factorization, Sparse Representation and principle component analysis are employed to describe the hidden semantic information in images, where they assume that the training and test sets are from same distribution. However, due to the considerable difference across the source and target domains result in environmental or device parameters, the traditional machine learning algorithms may fail. Transfer learning is a promising solution to deal with above problem, where the source and target data obey from different distributions. For enhancing the performance of model, transfer learning sends the knowledge from the source to target domain. Transfer learning benefits from sample reweighting of source data or feature projection of domains to reduce the divergence across domains. Sparse Coding joint with transfer learning has received more attention in many research fields, such as signal processing and machine learning where it makes the Representation more concise and easier to manipulate. Moreover, Sparse Coding facilitates an efficient content-based image indexing and retrieval. In this paper, we propose image classification via Sparse Representation and Subspace Alignment (SRSA) to deal with distribution mismatch across domains in low-level image Representation. Our approach is a novel image optimization algorithm based on the combination of instance-based and feature-based techniques. Under this framework, we reweight the source samples that are relevant to target samples using Sparse Representation. Then, we map the source and target data into their respective and independent subspaces. Moreover, we align the mapped subspaces to reduce the distribution mismatch across domains. The proposed approach is evaluated on various visual benchmark datasets with 14 experiments. Comprehensive experiments demonstrate that SRSA outperforms other latest machine learning and domain adaptation methods with significant difference.