• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권2호
• /
• pp.483-503
• /
• 2014

In this paper, human action recognition using pyramid histograms of oriented gradients and collaborative multi-task learning is proposed. First, we accumulate global activities and construct motion history image (MHI) for both RGB and depth channels respectively to encode the dynamics of one action in different modalities, and then different action descriptors are extracted from depth and RGB MHI to represent global textual and structural characteristics of these actions. Specially, average value in hierarchical block, GIST and pyramid histograms of oriented gradients descriptors are employed to represent human motion. To demonstrate the superiority of the proposed method, we evaluate them by KNN, SVM with linear and RBF kernels, SRC and CRC models on DHA dataset, the well-known dataset for human action recognition. Large scale experimental results show our descriptors are robust, stable and efficient, and outperform the state-of-the-art methods. In addition, we investigate the performance of our descriptors further by combining these descriptors on DHA dataset, and observe that the performances of combined descriptors are much better than just using only sole descriptor. With multimodal features, we also propose a collaborative multi-task learning method for model learning and inference based on transfer learning theory. The main contributions lie in four aspects: 1) the proposed encoding the scheme can filter the stationary part of human body and reduce noise interference; 2) different kind of features and models are assessed, and the neighbor gradients information and pyramid layers are very helpful for representing these actions; 3) The proposed model can fuse the features from different modalities regardless of the sensor types, the ranges of the value, and the dimensions of different features; 4) The latent common knowledge among different modalities can be discovered by transfer learning to boost the performance.

• Journal of Information Processing Systems
• /
• 제17권2호
• /
• pp.337-351
• /
• 2021

Automatically recognizing facial expressions in video sequences is a challenging task because there is little direct correlation between facial features and subjective emotions in video. To overcome the problem, a video facial expression recognition method using spatiotemporal recurrent neural network and feature fusion is proposed. Firstly, the video is preprocessed. Then, the double-layer cascade structure is used to detect a face in a video image. In addition, two deep convolutional neural networks are used to extract the time-domain and airspace facial features in the video. The spatial convolutional neural network is used to extract the spatial information features from each frame of the static expression images in the video. The temporal convolutional neural network is used to extract the dynamic information features from the optical flow information from multiple frames of expression images in the video. A multiplication fusion is performed with the spatiotemporal features learned by the two deep convolutional neural networks. Finally, the fused features are input to the support vector machine to realize the facial expression classification task. The experimental results on cNTERFACE, RML, and AFEW6.0 datasets show that the recognition rates obtained by the proposed method are as high as 88.67%, 70.32%, and 63.84%, respectively. Comparative experiments show that the proposed method obtains higher recognition accuracy than other recently reported methods.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권8호
• /
• pp.2101-2123
• /
• 2023

Recent studies have shown that the neural network-based binary code similarity detection technology performs well in vulnerability mining, plagiarism detection, and malicious code analysis. However, existing cross-architecture methods still suffer from insufficient feature characterization and low discrimination accuracy. To address these issues, this paper proposes a cross-architecture binary function similarity detection method based on composite feature model (SDCFM). Firstly, the binary function is converted into vector representation according to the proposed composite feature model, which is composed of instruction statistical features, control flow graph structural features, and application program interface calling behavioral features. Then, the composite features are embedded by the proposed hierarchical embedding network based on a graph neural network. In which, the block-level features and the function-level features are processed separately and finally fused into the embedding. In addition, to make the trained model more accurate and stable, our method utilizes the embeddings of predecessor nodes to modify the node embedding in the iterative updating process of the graph neural network. To assess the effectiveness of composite feature model, we contrast SDCFM with the state of art method on benchmark datasets. The experimental results show that SDCFM has good performance both on the area under the curve in the binary function similarity detection task and the vulnerable candidate function ranking in vulnerability search task.

• 한국정보과학회논문지:시스템및이론
• /
• 제31권5_6호
• /
• pp.347-352
• /
• 2004

본 논문에서는 메시지 패싱 시스템에서의 태스크 스케줄링에 대해 다룬다. 병렬/분산 시스템의 어플리케이션의 태스크에 대한 적절한 스케줄링이 이루어지지 않는 경우, 병렬/분산 처리를 이용한 이득을 기대하기는 어렵기 때문에 이 주제에 대한 연구는 컴퓨터 아키텍처의 발달과 함께 지속되고 있으며, 많은 연구들이 태스크 스케줄링에 대한 다양한 기법들을 제안하고 있다. 기존의 연구들은 공유 메모리 시스템을 가정하여 이루어졌기 때문에, 메시지 패싱 시스템에 기존의 기법을 적용하기가 힘들다. 본 논문에서는 기존 연구의 모델과 메시지 패싱 시스템의 통신 모델의 차이점으로 발생하는 통신비용의 누적을 고려하여 리스트 스케줄링 기법에 기초한 개선된 우선 순위 함수와 새로운 프로세서 선택 기준을 제안한다. 이들 두 가지 제안을 적용한 태스크 스케줄링 기법은 통신비용의 누적을 고려하지 않아 발생하는 비효율적인 스케줄링을 개선한다.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1986년도 한국자동제어학술회의논문집; 한국과학기술대학, 충남; 17-18 Oct. 1986
• /
• pp.518-526
• /
• 1986

This paper gives the detail of the work carried out to develop real-time control software for Flexible Manufacturing Systems. A basic design philosophy to implement such software is proposed. The major features are the partitioning of complicated control actions into simplified ones, structured programming and multi-threaded transaction-based tasks. The software operates on the basis of passing task-to-task messages via mailboxes, causing appropriate actions to be taken by each task. Each task represents a separate subprocess so that the subprocesses can be run simultaneously. The task-to-task message could be easily replaced by computer-to-computer communication, using LAN, demonstrating that the software methods developed produce a flexible designs for control software of an FMS. A method of linking such software to simulation software is suggested as a potentially powerful additional design-tool.

• 대한인간공학회지
• /
• 제14권1호
• /
• pp.25-35
• /
• 1995

For information processing tasks, it is an important cognitive skill to manipulate and store information, which is known as information intake. One of the tasks which greatly involve this skill would be a spreadsheet calculation task. In this study, a spreadsheet calculation task is analyzed by the cognitive task analysis based on the cognitive factors having been usef for a model of human information processing. By the results of the cognitive task analysis, the spreadsheet calculation tasks to be used in the experiments are designed and the testbattery of cognitive abilities assessment (CCAB ; complex cognitive asssessment battery) are selected. Then, the features of cognitive demands and a human performance model of the spreadsheet calculation task are suggested by means of correlation analysis, principal component factor analysis, and regression analysis of the results of the experiments on task performances and the assessment of cognitive abilities. Also, the application of the results of the study to job placement and further research issues are described.

• 대한인간공학회지
• /
• 제27권3호
• /
• pp.81-92
• /
• 2008

Musculoskeletal disorder (MSD) problems have been increasingly reported in hospital sector but the problems were not addressed with respect to holistic aspects of the target population in Korea. Often, it is required to understand how MSD symptoms are associated with factors such as personal, work environmental and psychosocial stressors. To examine features of association between sets of MSD symptoms and the factors, a questionnaire survey was conducted in a university hospital. A 140-item questionnaire was developed and used for collecting information including factors (e.g., job/occupation, task/activity, job stress) and MSD symptoms. A total of 1,091 workers (male 23.7% and female 76.3%) were finally determined for data analyses. Prevalence rate for the whole body was 72% and, among body parts, the highest was 48.7% for the shoulder, followed by 34.6%(the low back), 32.7%(the leg/foot), 27.9%(the neck), 26.7%(the wrist) and 12%(the elbow). The symptoms were significantly different by job/occupational variable in each of all body parts except the neck. The symptoms were very significantly different by task/activity variables in each of all body parts while those symptoms were significantly different by psychosocial variables, depending on body part and gender. In the logistic regression analyses performed for MSD symptoms by body part and each of 3 factors, odds ratio values varied, ranging from 0.7 to 3.3. The controls for reducing the symptoms were discussed on the basis of the findings. The results show that the MSD symptoms can remarkably vary by the factors and, in particular, can be highly differential for the task/activity factor. This study suggests that MSD symptom features be examined by using various factors and then a higher differential factor be primarily utilized for controling MSD symptoms in general industry including hospital settings.

• 로봇학회논문지
• /
• 제1권2호
• /
• pp.107-116
• /
• 2006

We propose a planning algorithm to automatically generate a robust behavior plan (RBP) with which mobile robots can achieve their task goal from any initial states under dynamically changing environments. For this, task description space (TDS) is formulated, where a redundant task configuration space and simulation model of physical space are employed. Successful task episodes are collected, where $A^*$ algorithm is employed. Interesting TDS state vectors are extracted, where occurrence frequency is used. Clusters of TDS state vectors are found by using state transition tuples and features of state transition tuples. From these operations, characteristics of successfully performed tasks by a simulator are abstracted and generalized. Then, a robust behavior plan is constructed as an ordered tree structure, where nodes of the tree are represented by attentive TDS state vector of each cluster. The validity of our method is tested by real robot's experimentation for a box-pushing-into-a-goal task.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.525-528
• /
• 2009

Recently, image classification has been an important task in various fields. Generally, the performance of image classification is not good without the adjustment of image features. Therefore, it is desired that the way of automatic feature extraction. In this paper, we propose an image classification method which adjusts image features automatically. We assume that texture features are useful in image classification tasks because natural images are composed of several types of texture. Thus, the classification accuracy rate is improved by using distribution of texture features. We obtain texture features by calculating image features from a current considering pixel and its neighborhood pixels. And we calculate image features from distribution of textures feature. Those image features are adjusted to image classification tasks using Genetic Algorithm. We apply proposed method to classifying images into "head" or "non-head" and "male" or "female".

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 Proceedings of the Korea Automatic Control Conference, 11th (KACC); Pohang, Korea; 24-26 Oct. 1996
• /
• pp.1-7
• /
• 1996

This paper addresses the control problem of a mobile robot supporting a task robot with needs to be positioned precisely. The main difficulty residing in the precise control of a mobile robot supporting a task robot is providing an accurate and stable base for the task robot. That is, the end-plate of the mobile robot which is the base of the task robot can not be positioned accurately without external position sensors. This difficulty is resolved in this paper through the vision information obtained from the camera attached at the end of a task robot. First of all, the camera parameters were measured by using the images of a fixed object captured by the camera. The measured parameters include the rotation, the position, the scale factor, and the focal length of the camera. These parameters could be measured by using the features of each vertex point for a hexagonal object and by using the pin-hole model of a camera. Using the measured pose(position and orientation) of the camera and the given kinematics of the task robot, we calculate a pose of the end-plate of the mobile robot, which is used for the precise control of the mobile robot. Experimental results for the pose estimations are shown.