• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.64 no.5
• /
• pp.766-778
• /
• 2015

In this paper, we design a hybrid system for recognition and tracking realized with the aid of polynomial based RBFNNs pattern classifier and particle filter. The RBFNN classifier is built by learning the training data for diverse pose images. The optimized parameters of RBFNN classifier are obtained by Particle Swarm Optimization(PSO). Testing data for pose image is used as a face image obtained under real situation, where the face image is detected by AdaBoost algorithm. In order to improve the recognition performance for a detected image, pose estimation as preprocessing step is carried out before the face recognition step. PCA is used for pose estimation, the pose of detected image is assigned for the built pose by considering the featured difference between the previously built pose image and the newly detected image. The recognition of detected image is performed through polynomial based RBFNN pattern classifier, and if the detected image is equal to target for tracking, the target will be traced by particle filter in real time. Moreover, when tracking is failed by PF, Adaboost algorithm detects facial area again, and the procedures of both the pose estimation and the image recognition are repeated as mentioned above. Finally, experimental results are compared and analyzed by using Honda/UCSD data known as benchmark DB.

• Journal of KIISE:Computing Practices and Letters
• /
• v.8 no.1
• /
• pp.107-115
• /
• 2002

We propose a new hybrid value predictor which achieves high performance by combining several predictors. Because the proposed hybrid value predictor can update the prediction table speculatively, it efficiently reduces the number of mispredicted instructions due to stale data. Also, the proposed predictor can enhance the prediction accuracy and efficiently decrease the hardware cost of predictor, because it allocates instructions into the best-suited predictor during instruction fetch stage by using the information of static classification which is obtained from the profile-based compiler implementation. For the 16-issue superscalar processors, simulation results based on the SimpleScalar/PISA tool set show that we achieve the average prediction rates of 73% by using speculative update and the average prediction rates of 88% by adding static classification for the SPECint95 benchmark programs.

• Journal of KIISE:Software and Applications
• /
• v.29 no.1_2
• /
• pp.1-9
• /
• 2002

Evolutionary artificial neural networks (EANNs) are towards the near optimal ANN using the global search of evolutionary instead of trial-and-error process. However, many real-world problems are too hard to be solved by only one ANN. Recently there has been plenty of interest on combining ANNs in the last generation to improve the performance and reliability. This paper proposes a new approach of constructing multiple ANNs which complement each other by speciation. Also, we develop a multiple ANN to combine the results in abstract, rank, and measurement levels. The experimental results on Australian credit approval data from UCI benchmark data set have shown that combining of the speciated EANNs have better recognition ability than EANNs which are not speciated, and the average error rate of 0.105 proves the superiority of the proposed EANNs.

• The KIPS Transactions:PartA
• /
• v.14A no.3 s.107
• /
• pp.167-172
• /
• 2007

VM(Virtual Machine) can be considered as a software processor which interprets the machine code. Also, it is considered as a conceptional computer that consists of logical system configuration. But, the execution speed of VM system is much slower than that of a real processor system. So, it is very important to optimize the code for virtual machine to enhance the execution time. Especially the optimizer for a virtual machine code on embedded devices requires the highly efficient performance to the ordinary optimizer in the respect to the optimized ratio about cost. Fundamentally, functions and basic blocks which influence the execution time of virtual machine is found, and then an optimization for them nay get the high efficiency. In this paper, we designed and implemented the optimizer for the virtual(or abstract) machine code(VMC) using profiling. Firstly, we defined the profiling information which is necessary to the optimization of VMC. The information can be obtained from dynamically executing the machine code. And we implemented VMC optimizer using the profiling information. In our implementation, the VMC is SIL(Standard Intermediate Language) that is an intermediate code of EVM(Embedded Virtual Machine). Also, we tried a benchmark test for the VMC optimizer and obtained reasonable results.

• Smart Structures and Systems
• /
• v.24 no.5
• /
• pp.631-639
• /
• 2019

Finite element analysis is one of the important methods to study the structural performance. Due to the simplification, discretization and error of structural parameters, numerical model errors always exist. Besides, structural characteristics may also change because of material aging, structural damage, etc., making the initial finite element model cannot simulate the operational response of the structure accurately. Based on Bayesian methods, the initial model can be updated to obtain a more accurate numerical model. This paper presents the work on the field test, modal identification and model updating of a Chinese reinforced concrete pagoda. Based on the ambient vibration test, the acceleration response of the structure under operational environment was collected. The first six translational modes of the structure were identified by the enhanced frequency domain decomposition method. The initial finite element model of the pagoda was established, and the elastic modulus of columns, beams and slabs were selected as model parameters to be updated. Assuming the error between the measured mode and the calculated one follows a Gaussian distribution, the posterior probability density function (PDF) of the parameter to be updated is obtained and the uncertainty is quantitatively evaluated based on the Bayesian statistical theory and the Metropolis-Hastings algorithm, and then the optimal values of model parameters can be obtained. The results show that the difference between the calculated frequency of the finite element model and the measured one is reduced, and the modal correlation of the mode shape is improved. The updated numerical model can be used to evaluate the safety of the structure as a benchmark model for structural health monitoring (SHM).

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.9
• /
• pp.379-390
• /
• 2019

In this paper, we propose a novel deep neural network model for Vision-and-Language Navigation (VLN) named LVLN (Landmark-based VLN). In addition to both visual features extracted from input images and linguistic features extracted from the natural language instructions, this model makes use of information about places and landmark objects detected from images. The model also applies a context-based attention mechanism in order to associate each entity mentioned in the instruction, the corresponding region of interest (ROI) in the image, and the corresponding place and landmark object detected from the image with each other. Moreover, in order to improve the success rate of arriving the target goal, the model adopts a progress monitor module for checking substantial approach to the target goal. Conducting experiments with the Matterport3D simulator and the Room-to-Room (R2R) benchmark dataset, we demonstrate high performance of the proposed model.

• International Journal of Computer Science & Network Security
• /
• v.21 no.8
• /
• pp.288-296
• /
• 2021

Automated face recognition in a runtime environment is gaining more and more important in the fields of surveillance and urban security. This is a difficult task keeping in mind the constantly volatile image landscape with varying features and attributes. For a system to be beneficial in industrial settings, it is pertinent that its efficiency isn't compromised when running on roads, intersections, and busy streets. However, recognition in such uncontrolled circumstances is a major problem in real-life applications. In this paper, the main problem of face recognition in which full face is not visible (Occlusion). This is a common occurrence as any person can change his features by wearing a scarf, sunglass or by merely growing a mustache or beard. Such types of discrepancies in facial appearance are frequently stumbled upon in an uncontrolled circumstance and possibly will be a reason to the security systems which are based upon face recognition. These types of variations are very common in a real-life environment. It has been analyzed that it has been studied less in literature but now researchers have a major focus on this type of variation. Existing state-of-the-art techniques suffer from several limitations. Most significant amongst them are low level of usability and poor response time in case of any calamity. In this paper, an improved face recognition system is developed to solve the problem of occlusion known as FRS-OCC. To build the FRS-OCC system, the color and texture features are used and then an incremental learning algorithm (Learn++) to select more informative features. Afterward, the trained stack-based autoencoder (SAE) deep learning algorithm is used to recognize a human face. Overall, the FRS-OCC system is used to introduce such algorithms which enhance the response time to guarantee a benchmark quality of service in any situation. To test and evaluate the performance of the proposed FRS-OCC system, the AR face dataset is utilized. On average, the FRS-OCC system is outperformed and achieved SE of 98.82%, SP of 98.49%, AC of 98.76% and AUC of 0.9995 compared to other state-of-the-art methods. The obtained results indicate that the FRS-OCC system can be used in any surveillance application.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.1
• /
• pp.13-19
• /
• 2019

The accuracy of machine learning is greatly affected by amount of learning data and quality of data. Collecting existing Web-based learning data has danger that data unrelated to actual learning can be collected, and it is impossible to secure data transparency. In this paper, we propose a method for collecting data directly in parallel by blocks in a block - chain structure, and comparing the data collected by each block with data in other blocks to select only good data. In the proposed system, each block shares data with each other through a chain of blocks, utilizes the All-reduce structure of Parallel-SGD to select only good quality data through comparison with other block data to construct a learning data set. Also, in order to verify the performance of the proposed architecture, we verify that the original image is only good data among the modulated images using the existing benchmark data set.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.9
• /
• pp.1049-1059
• /
• 2019

We propose a learning-based image segmentation algorithm. Starting from super-pixels, our method learns the probability of merging two regions based on the ground truth made by humans. The learned information is used in determining whether the two regions should be merged or not in a segmentation stage. Unlike exiting learning-based algorithms, we use both local and object information. The local information represents features computed from super-pixels and the object information represent high level information available only in the learning process. The object information is considered as privileged information, and we can use a framework that utilize the privileged information such as SVM+. In experiments on the Berkeley Segmentation Dataset and Benchmark (BSDS 500) and PASCAL Visual Object Classes Challenge (VOC 2012) data set, out model exhibited the best performance with a relatively small training data set and also showed competitive results with a sufficiently large training data set.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3729-3749
• /
• 2021

At present, deep convolution network-based salient object detection (SOD) has achieved impressive performance. However, it is still a challenging problem to make full use of the multi-scale information of the extracted features and which appropriate feature fusion method is adopted to process feature mapping. In this paper, we propose a new adjacency auxiliary network (AANet) based on multi-scale feature fusion for SOD. Firstly, we design the parallel connection feature enhancement module (PFEM) for each layer of feature extraction, which improves the feature density by connecting different dilated convolution branches in parallel, and add channel attention flow to fully extract the context information of features. Then the adjacent layer features with close degree of abstraction but different characteristic properties are fused through the adjacent auxiliary module (AAM) to eliminate the ambiguity and noise of the features. Besides, in order to refine the features effectively to get more accurate object boundaries, we design adjacency decoder (AAM_D) based on adjacency auxiliary module (AAM), which concatenates the features of adjacent layers, extracts their spatial attention, and then combines them with the output of AAM. The outputs of AAM_D features with semantic information and spatial detail obtained from each feature are used as salient prediction maps for multi-level feature joint supervising. Experiment results on six benchmark SOD datasets demonstrate that the proposed method outperforms similar previous methods.