یاماناشی، کوفو، ژاپن
یاماناشی، کوفو، ژاپن
شهید بهشتی، تهران، ایران
داده ای یافت نشد
Long term temporal representation methods demand high computational cost, restricting their practical use in real world applications. We propose a two-step deep residual method for efficiently learning long-term discriminative temporal representation, whilst significantly reducing computational cost. In the first step, a novel self-supervision deep temporal embedding method is presented to embed repetitive short-term motions at a cluster-friendly feature space. In the second step, an efficient temporal representation is made by leveraging the differences between the original data and its associated repetitive motion clusters as a novel deep residual method. Experimental results demonstrate that, the proposed method achieves competitive res
We address the following problem: Given a simple polygon P with n vertices and two points s and t inside it, find a minimum link path between them such that a given target point q is visible from at least one point on the path. The method is based on partitioning a portion of P into a number of faces of equal link distance from a source point. This partitioning is essentially a shortest path map (SPM). In this paper, we present an optimal algorithm with O (n) time bound, which is the same as the time complexity of the standard minimum link paths problem.
The purpose of this note is to give a simple proof for a necessary and sufficient condition for visibility paths in simple polygons. A visibility path is a curve such that every point in the simple polygon is visible from at least one point of the path. This result is required for the shortest watchman route problem specially when the route is restricted to curved paths.
We study the query version of constrained minimum link paths between two points inside a simple polygon with vertices such that there is at least one point on the path, visible from a query point. Initially, we solve this problem for two given points , and a query point . Then, the proposed solution would be extended to a general case for three arbitrary query points , and . In the former, we propose an algorithm with preprocessing time. Extending this approach for the latter case, we develop an algorithm with preprocessing time. The link distance between , with visibility , and its path are provided in time and for the above two cases, where is the number of links.
We investigate a practical variant of the well-known polygonal visibility path (watchman) problem. For a polygon $ P $, a minimum link visibility path is a polygonal visibility path in $ P $ that has the minimum number of links. The problem of finding a minimum link visibility path is NP-hard for simple polygons. If the link-length (number of links) of a minimum link visibility path (tour) is $ OPT $ for a simple polygon $ P $ with $ n $ vertices and $ k $ nonredundant cuts, we provide an algorithm with $ O (kn^ 2) $ runtime that produces polygonal visibility paths (or tours) of link-length at most $(\gamma+ a_l/(k-1)) OPT $(or $(\gamma+ a_l/k) OPT $), where $ a_l $ is an output sensitive parameter and $\gamma $ is the approximation factor
Deep networks have recently achieved great success in feature learning problem on various computer vision applications. Among different approaches in deep learning, unsupervised methods have attracted a lot of attention particularly to problems with limited training data. However, compared with supervised methods, unsupervised deep learning methods usually suffer from lower accuracy and higher computational time. To deal with these problems, we aim to restrict the number of connections between successive layers while enhancing discriminatory power using a data-driven approach. To this end, we propose a novel deep multi-view ensemble model. The structure of each layer is composed of an ensemble of encoders or decoders and mask operations. Th
Visual stimulus decoding is an increasingly important challenge in neuroscience. The goal is to classify the activity patterns from the human brain; during the sighting of visual objects. One of the crucial problems in the brain decoder is the selecting informative voxels. We propose a meta‐heuristic voxel selection framework for brain decoding. It is composed of four phases: preprocessing of fMRI data; filtering insignificant voxels; postprocessing; and meta‐heuristics selection. The main contribution is benefiting a meta‐heuristics search algorithm to guide a wrapper voxel selection. The main criterion to nominate a voxel is based on its mutual information with the provided stimulus label. The results show impressive accuracy rates
The heterogeneity of a network causes major challenges for link prediction in heterogeneous complex networks. To deal with this problem, supervised link prediction could be applied to integrate heterogeneous features extracted from different nodes/relations. However, supervised link prediction might be faced with highly imbalanced data issues which results in undesirable false prediction rate. In this paper, we propose a new kernel-based one-class link predictor in heterogeneous complex networks. Assuming a set of available meta-paths, a graph kernel is extracted based on each meta-path. Then, they are combined to form a single kernel function. Afterwards, one class support vector machine (OC-SVM) would be applied on the positi
Finding similar time series has attracted a lot of interest and much research has been done recently as a result . For the reason of high dimension of the time series data, finding a good answer to this problem is difficult. Encounter with these high dimensional data requires us to use dimension reduction techniques, and then performing data mining tasks on reduced dataset. Several time series dimension reduction techniques have been proposed before, such as DFT , DWT , SVD , PAA  and , APCA , PLA , SAX  and many others, but we cannot simply choose an arbitrary compression algorithm . Each one of these algorithms has different answers to a unique problem. The main contribution of this paper reviewing the time
The majority of the link prediction measures in heterogeneous complex networks rely on the nodes connectivities while less attention has been paid to the importance of the nodes and paths. In this paper, we propose some new meta-path based statistical similarity measures to properly perform link prediction task. The main idea in the proposed measures is to drive some co-occurrence events in a number of co-occurrence matrices that are occurred between the visited nodes obeying a meta-path. The extracted co-occurrence matrices are analyzed in terms of the energy, inertia, local homogeneity, correlation, and information measure of correlation to determine various information theoretic measures. We evaluate the proposed measures, denoted as lin
Hidden Markov model (HMM) has been widely applied in human action recognition. In this paper an extension of HMM called fuzzy hidden Markov model (fuzzy HMM) is used for action recognition. It tries to increase the classification performance and decrease the information loss due to feature vector quantisation. Using fuzzy concepts with HMM leads to better recognition of similar actions such as walking, jogging and running. Two feature extraction methods including skeleton and space-time approaches are used for action representation. Actions could be represented efficiently using skeleton features where scene background is plain. Space-time features are extracted directly from video, and therefore avoid possible failures of other pre-process
Large scale distributed systems employ thousands of resources which inevitably suffer from the unavailability issue. Serious side effects like unexpected delay or failure in the application execution are probable in case of such an issue. The imposed outcome might then be catastrophic consequences for real time applications or penalties for the service providers. Better prediction of the resource unavailability helps diminishing the undesired outcomes. This paper proposes a resource availability prediction algorithm for the mentioned goal. The resource availability variation is modeled as a stochastic process. By analyzing the availability information of NDU resources and both physical and virtual machines of the PlantLab, we found that the
We add a new phase, called reforming phase, to support vector data description (SVDD) between the training and testing phases. The reforming phase enables us to reconsider the SVDD’s assumption of the uniformity of features in calculating the distance of an object to the center of hypersphere. In the reforming phase, the features are assumed as a group of experts who have different impacts in overall outlier detection. In doing so, the proportion of each feature in the distance of an object to the center of hypersphere is specified. Subsequently, the opinions of the experts about the label of the corresponding object are determined based on these measured proportions. By using different group decision-making methods for aggre
A study on one of the most important issues in a human action recognition task, i.e. how to create proper data representations with a high-level abstraction from large dimensional noisy video data, is carried out. Most of the recent successful studies in this area are mainly focused on deep learning. Deep learning methods have gained superiority to other approaches in the field of image recognition. In this survey, the authors first investigate the role of deep learning in both image and video processing and recognition. Owing to the variety and plenty of deep learning methods, the authors discuss them in a comparative form. For this purpose, the authors present an analytical framework to classify and to evaluate these methods based on some
Recently, a number of meta-path based similarity indices like PathSim, HeteSim, and random walk have been proposed for link prediction in heterogeneous complex networks. However, these indices suffer from two major drawbacks. Firstly, they are primarily dependent on the connectivity degrees of node pairs without considering the further information provided by the given meta-path. Secondly, most of them are required to use a single and usually symmetric meta-path in advance. Hence, employing a set of different meta-paths is not straightforward. To tackle with these problems, we propose a mutual information model for link prediction in heterogeneous complex networks. The proposed model, called as Meta-path based Mutual Information Index (MMI)
Many real-world phenomena can be intelligently modeled with complex networks. In this study, the most important optimization issues in complex networks are reviewed including network modeling, network sampling, network partitioning, link prediction, and influence maximization. For each issue, some conventional methods are firstly studied. Then, the most important challenges and future directions would be discussed.
Social Networks (SNs) have gained a lot of popularity on the Internet and become a hot research topic attracting many professionals from diverse areas. Recently, Location-based Social Networks (LBSNs) have attracted millions of users, experiencing a huge popularity increase over a short period of time. In Location-based social network, users can easily set their locations as a new interactive way to share with friends to inform them of their current location. In this paper, we explores the community structure of a location based social network data and propose a new link predictor for its. In the proposed approach, the network is firstly partitioned into a number of groups. Then, a supervised link predictor is learnt for each group. To do t
Emotion is expressed via facial muscle movements, speech, body and hand gestures, and various biological signals like heart beating. However, the most natural way that humans display emotion is facial expression. Facial expression recognition is a great challenge in the area of computer vision for the last two decades. This paper focuses on facial expression to identify seven universal human emotions ie anger, disgust, fear, happiness, sadness, surprise, and neu7tral. Unlike the majority of other approaches which use the whole face or interested regions of face, we restrict our facial emotion recognition (FER) method to analyze human emotional states based on eye region changes. The reason of using this region is that eye region is one of t
اطلاعیه ای درج نشده است