Search

Deep neural networks are extensively used in text mining and Natural Language Processing is to enable computers to understand, analyze, and generate natural language data, such as text or speech, but semantic resources, such as taxonomies and ontologies, are not fully included in deep learning. In this paper, we use Deep Convolutional Neural Network (Deep CNN) to classify research papers using the Computer Science Ontology, an ontology of research areas in the field of computer science. It takes as input the abstract and keywords of a particular research paper and returns the relevant research topic. To evaluate our ontology, we used a gold standard dataset that includes research articles. To further improve text classification results, we propose to design a Deep CNN model. We then used ontology matching to reduce the classes and get better results. Experimental results show that the proposed approach outperforms the one with the highest precision, recall, and F1-score. [ABSTRACT FROM AUTHOR]

Finding the saddle point of a matrix game is a classical problem that arises in various fields, e.g., economics, computer science, and engineering. The standard problem-solving methods consist of formulating the problem as a linear program (LP). However, this approach seems less efficient when many instances need to be solved. In this paper, we propose a Convolutional Neural Network based approach, which is able to predict both the strategy profile (x , y) and the optimal value v of the game. We call this approach Matrix Game-Conventional Neural Network or MG-CNN for short. Thanks to a global pooling technique, MG-CNN can solve matrix games with different shapes. We propose a specialized algorithm to train MG-CNN, which includes both data generation and model training. Our numerical experiments show that MG-CNN outperforms standard LP solvers in terms of computational CPU time and provides a high-quality prediction. [ABSTRACT FROM AUTHOR]

Vehicle Re-Identification (Re-ID) is a challenging vision task mainly because the appearance of a vehicle varies dramatically under different viewpoints. Moreover, different vehicles with the same model and color commonly show similar appearance, thus are hard to be distinguished. To alleviate negative effects of viewpoint variance, we design a multi-view branch network where each branch learns a viewpoint-specific feature without parameter sharing. Being able to focus on a limited range of viewpoints, this viewpoint-specific feature performs substantially better than the general feature learned by an uniform network. To further differentiate visually similar vehicles, we strengthen the discriminative power on their subtle local differences by introducing a spatial attention model into each feature learning branch. The multi-view feature learning and spatial attention learning compose our neural network architecture, which is trained end to end with the softmax loss and triplet loss, respectively. We evaluate our methods on two large vehicle Re-ID datasets, i.e., VehicleID and VeRi-776, respectively. Extensive experiments show that our methods achieve promising performance. For example, we achieve mAP accuracy of 76.78% and 72.53% on VehicleID and VeRi-776 dataset respectively, substantially better than current state-of-the art. [ABSTRACT FROM AUTHOR]

In order to improve the robustness of text detector on scene text of various scales, a single shot text detector that combines local and non-local features is proposed in this paper. A dilated inception module for local feature extraction and a text self-attention module for non-local feature extraction are presented, and these two kinds of modules are integrated into single shot detector (SSD) of generic object detection so as to perform multi-oriented text detection in natural scene. The proposed modules make a contribution to richer and wider receptive field and enhance feature representation. Furthermore, the performance of our text detector is improved. In addition, compared with previous text detectors based on SSD which classify positive and negative samples depending on default boxes, we exploit pixels as reference for more accurate matching with ground truth which avoids complex anchor design. Furthermore, to evaluate the effectiveness of the proposed method, we carry out several comparative experiments on public standard benchmarks and analyze the experimental results in detail. The experimental results illustrate that the proposed text detector can compete with the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

This paper presents two on-going efforts of the Pacific Earthquake Engineering Research (PEER) center in the area of structural health monitoring. The first is data-driven damage assessment, which focuses on using data from instrumented buildings to compute the values of damage features. Using machine learning algorithms, these damage features are used for rapid identification of the level and location of damage after earthquakes. One of the damage features identified to be highly efficient is the cumulative absolute velocity. The second is vision-based automated damage identification and assessment from images. Deep learning techniques are used to conduct several identification tasks from images, examples of which are the structural component type, and level and type of damage. The objective is to use crowdsourcing, allowing the general public to take photographs of damage and upload them to a server where damage is automatically identified using deep learning algorithms. The paper also introduces PEER's effort and preliminary results in engaging the engineering and computer science communities in such developments through the PEER Hub Image- Net (Φ-Net) challenge. [ABSTRACT FROM AUTHOR]

Convolutional Neural Networks (CNNs) have been applied to visual tracking with demonstrated success in recent years. Most CNN-based trackers utilize hierarchical features extracted from a certain layer to represent the target. However, features from a certain layer are not always effective for distinguishing the target object from the backgrounds especially in the presence of complicated interfering factors (e.g., heavy occlusion, background clutter, illumination variation, and shape deformation). In this work, we propose a CNN-based tracking algorithm which hedges deep features from different CNN layers to better distinguish target objects and background clutters. Correlation filters are applied to feature maps of each CNN layer to construct a weak tracker, and all weak trackers are hedged into a strong one. For robust visual tracking, we propose a hedge method to adaptively determine weights of weak classifiers by considering both the difference between the historical as well as instantaneous performance, and the difference among all weak trackers over time. In addition, we design a Siamese network to define the loss of each weak tracker for the proposed hedge method. Extensive experiments on large benchmark datasets demonstrate the effectiveness of the proposed algorithm against the state-of-the-art tracking methods. [ABSTRACT FROM AUTHOR]

Depth information has been discovered to improve the performance of computer vision tasks, such as semantic segmentation and object recognition. However, careful acquisition of depth data needs highly developed depth sensors which are expensive. As a classic computer vision task, depth estimation from a single image has obtained promising results based on supervised learning methods. In this paper, we investigate the extension of color images with corresponding deep-regressed depth images in boosting the performance of semantic segmentation. Furthermore, the usage of combining color channels with the estimated depth or the ground truth depth channel is compared. Specifically, there are two stages in our work. Firstly, we adopt the framework of convolutional neural networks (CNN) for the depth estimation by combing the global depth network and the depth gradient network. After refining based on these two networks, the depth image map can be estimated in a deep-regressed manner. Secondly, after augmenting the color images with the predicted depth images, fully convolutional networks (FCN) are further used to implement the pixel-level semantic labeling. In the experiments, we employ two popular RGBD datasets, i.e., SUNRGBD and NYUDv2, for 37 and 40-class semantic segmentation, respectively. By comparing with the ground truth depth images, experimental results demonstrate that the networks trained on the estimated depth images can achieve comparable performance on facilitating the accuracy of semantic segmentation task. [ABSTRACT FROM AUTHOR]

Different words in discourse arguments usually have varying contributions on the recognition of implicit discourse relations. Following this intuition, we propose two attention-based neural networks, namely inner attention model and outer attention model , to learn better discourse representation by automatically estimating the degrees of relevance of words to discourse relations. The former model only utilizes the information inside discourse arguments, while the latter model builds upon an outside semantic memory to exploit general world knowledge. Both models are capable of assigning more weights to relation-relevant words, and operate in an end-to-end manner. Upon these two models, we further propose a full attention model that combines their strengths into a unified framework. Extensive experiments on the PDTB data set show that our model significantly benefits from highlighting relation-relevant words and yields competitive and even better results against several state-of-the-art systems. [ABSTRACT FROM AUTHOR]

This paper describes an efficient generation of large-scale dataset of human depth images with body part labels. The size of image datasets has recently been increasingly important as it is shown to be strongly related to the performance of learning-based classifiers. In human pose recognition, many datasets for ordinary poses like standing, walking, and doing gestures have already been developed and effectively utilized. However, those for unusual ones like lying fainted and crouching do not exist. Pose recognition for such cases may have a large potential applicability to various assistive scenarios. Moreover, locating each body part could also be important for an accurate care and diagnosis or anomaly detection. We therefore develop a method of generating body part-annotated depth images in various body shapes and poses, which are handled by a flexible human body model and a motion capture system, respectively. We constructed a dataset of 10,076 images with eight body types for various sitting poses. The effectiveness of generated dataset is verified by part labeling tasks with a fully convolutional network (FCN) for synthetic and real test data. [ABSTRACT FROM AUTHOR]

Bag of Words (BoW) model and Convolutional Neural Network (CNN) are two milestones in visual recognition. Both BoW and CNN require a feature pooling operation for constructing the frameworks. Particularly, the max-pooling has been validated as an efficient and effective pooling method compared with other methods such as average pooling and stochastic pooling. In this paper, we first evaluate different pooling methods, and then propose a new feature pooling method termed as selective, discriminative and equalizing pooling (SDE). The SDE representation is a feature learning mechanism by jointly optimizing the pooled representations with the target of learning more selective, discriminative and equalizing features. We use bilevel optimization to solve the joint optimization problem. Experiments on seven benchmark datasets (including both single-label and multi-label ones) well validate the effectiveness of our framework. Particularly, we achieve the state-of-the-art fused results (mAP) of 93.21 and 93.97% on the PASCAL VOC2007 and VOC2012 datasets, respectively. [ABSTRACT FROM AUTHOR]

Computer vision is one of the most frontier technologies in computer science. It is used to build artificial systems to extract valuable information from images and has a broad range of applications in various areas such as agriculture, business, and healthcare. Convolutional neural networks represent the key algorithms in computer vision, and in recent years, they have attained notable advances in many real-world problems. The accuracy of the network for a particular task profoundly relies on the hyperparameters' configuration. Obtaining the right set of hyperparameters is a time-consuming process and requires expertise. To approach this concern, we propose an automatic method for hyperparameters' optimization and structure design by implementing enhanced metaheuristic algorithms. The aim of this paper is twofold. First, we propose enhanced versions of the tree growth and firefly algorithms that improve the original implementations. Second, we adopt the proposed enhanced algorithms for hyperparameters' optimization. First, the modified metaheuristics are evaluated on standard unconstrained benchmark functions and compared to the original algorithms. Afterward, the improved algorithms are employed for the network design. The experiments are carried out on the famous image classification benchmark dataset, the MNIST dataset, and comparative analysis with other outstanding approaches that were tested on the same problem is conducted. The experimental results show that both proposed improved methods establish higher performance than the other existing techniques in terms of classification accuracy and the use of computational resources. [ABSTRACT FROM AUTHOR]