• IEIE Transactions on Smart Processing and Computing
• /
• v.6 no.1
• /
• pp.53-59
• /
• 2017

Optical character recognition (OCR) automatically recognizes text in an image. OCR is still a challenging problem in computer vision. A successful solution to OCR has important device applications, such as text-to-speech conversion and automatic document classification. In this work, we analyze character recognition performance using the current state-of-the-art deep-learning structures. One is the AlexNet structure, another is the LeNet structure, and the other one is the SPNet structure. For this, we have built our own dataset that contains digits and upper- and lower-case characters. We experiment in the presence of salt-and-pepper noise or Gaussian noise, and report the performance comparison in terms of recognition error. Experimental results indicate by five-fold cross-validation that the SPNet structure (our approach) outperforms AlexNet and LeNet in recognition error.

• KIEAE Journal
• /
• v.14 no.2
• /
• pp.21-28
• /
• 2014

This paper provides the up to date review of the shading effects on PV module performance and the associated detection algorithm related to the maximum power point tracking. It includes the brief explanations of the MMP variations due to the shading occurrence on the PV modules. Review of experimental and simulation studies highlighting the significant impacts of shading on PV efficiencies were presented. The literature indicates that even the partial shading of a single cell can greatly drop the entire module voltage and power efficiency. The MMP tracking approaches were also reviewed in this study. Both conventional and advanced soft computing methods such as ANN, FLC and EA were described for the proper tracking of MMP under shaded conditions. This paper would be the basic source and the comprehensive information associated with the shading effects and relevant MPP tracking technique.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.3
• /
• pp.827-833
• /
• 2016

In this paper, we propose a deep least squares support vector machine (LS-SVM) for regression problems, which consists of the input layer and the hidden layer. In the hidden layer, LS-SVMs are trained with the original input variables and the perturbed responses. For the final output, the main LS-SVM is trained with the outputs from LS-SVMs of the hidden layer as input variables and the original responses. In contrast to the multilayer neural network (MNN), LS-SVMs in the deep LS-SVM are trained to minimize the penalized objective function. Thus, the learning dynamics of the deep LS-SVM are entirely different from MNN in which all weights and biases are trained to minimize one final error function. When compared to MNN approaches, the deep LS-SVM does not make use of any combination weights, but trains all LS-SVMs in the architecture. Experimental results from real datasets illustrate that the deep LS-SVM significantly outperforms state of the art machine learning methods on regression problems.

• ETRI Journal
• /
• v.38 no.2
• /
• pp.347-355
• /
• 2016

In face examinations, gender classification (GC) is one of several fundamental tasks. Recent literature on GC primarily utilizes datasets containing high-resolution images of faces captured in uncontrolled real-world settings. In contrast, there have been few efforts that focus on utilizing low-resolution images of faces in GC. We propose a GC method based on a pixel classifier boosting with modified census transform features. Experiments are conducted using large datasets, such as Labeled Faces in the Wild and The Images of Groups, and standard protocols of GC communities. Experimental results show that, despite using low-resolution facial images that have a 15-pixel inter-ocular distance, the proposed method records a higher classification rate compared to current state-of-the-art GC algorithms.

• ETRI Journal
• /
• v.31 no.1
• /
• pp.10-20
• /
• 2009

In this paper, a video broadcasting system between a home-server-type device and a mobile device is proposed. The home-server-type device can automatically extract semantic information from video contents, such as news, a soccer match, and a baseball game. The indexing results are utilized to convert the original video contents to a digested or arranged format. From the mobile device, a user can make recording requests to the home-server-type devices and can then watch and navigate recorded video contents in a digested form. The novelty of this study is the actual implementation of the proposed system by combining the actual IT environment that is available with indexing algorithms. The implementation of the system is demonstrated along with experimental results of the automatic video indexing algorithms. The overall performance of the developed system is compared with existing state-of-the-art personal video recording products.

• ETRI Journal
• /
• v.39 no.3
• /
• pp.428-436
• /
• 2017

State-of-the-art processors require increasingly complicated memory services for high performance and low power consumption. In particular, they request transfers within a burst in a wrap-around order to minimize the miss penalty of a cache. However, synchronous dynamic random access memories (SDRAMs) do not always generate transfers in the wrap-round order required by the processors. Thus, a memory subsystem rearranges the SDRAM transfers in the wrap-around order, but the rearrangement process may increase memory latency and waste the bandwidth of on-chip interconnects. In this paper, we present a memory subsystem that is effective for the wrapping bursts of a cache. The proposed memory subsystem makes SDRAMs generate transfers in an intermediate order, where the transfers are rearranged in the wrap-around order with minimal penalties. Then, the transfers are delivered with priority, depending on the program locality in space. Experimental results showed that the proposed memory subsystem minimizes the memory performance loss resulting from wrapping bursts and, thus, improves program execution time.

• Journal of computational fluids engineering
• /
• v.21 no.1
• /
• pp.1-9
• /
• 2016

Atmospheric icing may have significant effects not only on safety of aircraft in air, but also on performance of wind turbine and power networks on ground. Thus, ice protection measure should be developed to protect these systems from icing hazards. A very efficient method is the electro-thermal de-icing based on a process by which ice accretion is melted and blown away through aerodynamic forces. In this computational study, a state-of-the-art icing code, FENSAP-ICE, was used for the analysis of electro thermal de-icing system. Computational results including detailed conjugate heat transfer analysis were then validated with experimental data. Further, the computational model was applied to the DU21 airfoil section of NREL 5MW wind turbine with calculated heater parameters.

• Proceedings of the KSME Conference
• /
• 2002.08a
• /
• pp.419-422
• /
• 2002

A three-dimensional inlet flow structure inside a microfluidic element has been investigated using a micro-PIV(particle image velocimetry) measurement as well as a numerical analysis. The present study employs a state-of-art micro-PIV system which consists of epi-fluorescence microscope, 620nm diameter fluorescent seed particles and an 8-bit megapixel CCD camera. For the numerical analysis, a commercial software CFD-ACE+(V6.6) was employed for comparison with experimental data. Fixed pressure boundary condition and a 39900 structured grid system was used for numerical analysis. Velocity vector fields with a resolution of $6.7{\times}6.7{\mu}m$ has been obtained, and the attention has been paid on the effect of varying measurement conditions of particle diameter and particle concentration on the resulting PIV results. In this study, the microfluidic elements were fabricated on plastic chips by means of MEMS processes and a subsequent melding process.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.87-90
• /
• 2005

Today's state-of the-art speech recognition systems typically use continuous distribution hidden Markov models with the mixtures of Gaussian distributions. To obtain higher recognition accuracy, the hidden Markov models typically require huge number of Gaussian distributions. Such speech recognition systems have problems that they require too much memory to run, and are too slow for large applications. Many approaches are proposed for the design of compact acoustic models. One of those models is subspace distribution clustering hidden Markov model. Subspace distribution clustering hidden Markov model can represent original full-space distributions as some combinations of a small number of subspace distribution codebooks. Therefore, how to make the codebook is an important issue in this approach. In this paper, we report some experimental results on various quantization methods to make more accurate models.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.6
• /
• pp.2310-2332
• /
• 2020

In recent years, deep learning techniques have achieved tremendous successes in natural language processing, speech recognition and image processing. Collaborative filtering(CF) recommendation is one of widely used methods and has significant effects in implementing the new recommendation function, but it also has limitations in dealing with the problem of poor scalability, cold start and data sparsity, etc. Combining the traditional recommendation algorithm with the deep learning model has brought great opportunity for the construction of a new recommender system. In this paper, we propose a novel collaborative recommendation model based on auxiliary stacked denoising autoencoder(ASDAE), the model learns effective the preferences of users from auxiliary information. Firstly, we integrate auxiliary information with rating information. Then, we design a stacked denoising autoencoder based collaborative recommendation model to learn the preferences of users from auxiliary information and rating information. Finally, we conduct comprehensive experiments on three real datasets to compare our proposed model with state-of-the-art methods. Experimental results demonstrate that our proposed model is superior to other recommendation methods.