• Journal of Information Processing Systems
• /
• v.18 no.5
• /
• pp.701-710
• /
• 2022

With the development of artificial intelligence technology, various methods have been developed for recognizing objects in images using machine learning. Image segmentation is the most effective among these methods for recognizing objects within an image. Conventionally, image datasets of various classes are trained simultaneously. In situations where several classes require segmentation, all datasets have to be trained thoroughly. Such repeated training results in low training efficiency because most of the classes have already been trained. In addition, the number of classes that appear in the datasets affects training. Some classes appear in datasets in remarkably smaller numbers than others, and hence, the training errors will not be properly reflected when all the classes are trained simultaneously. Therefore, a new method that separates some classes from the dataset is proposed to improve efficiency during training. In addition, the accuracies of the conventional and proposed methods are compared.

• Journal of IKEEE
• /
• v.20 no.4
• /
• pp.351-360
• /
• 2016

Recently, machine learning algorithms, especially deep learning-based algorithms, have been receiving attention due to its high classification performance. Among the algorithms, Convolutional Neural Network(CNN) is known to be efficient for image processing tasks used for Advanced Driver Assistance Systems(ADAS). However, it is difficult to achieve real-time processing for CNN in vehicle embedded software environment due to the repeated operations contained in each layer of CNN. In this paper, we propose a hardware accelerator which enhances the execution time of CNN by parallelizing the repeated operations such as convolution. Xilinx ZC706 evaluation board is used to verify the performance of the proposed accelerator. For $36{\times}36$ input images, the hardware execution time of CNN is 2.812ms in 100MHz clock frequency and shows that our hardware can be executed in real-time.

• Korean Journal of Cognitive Science
• /
• v.21 no.4
• /
• pp.603-625
• /
• 2010

Neural representation of a visual object is distributed across visual cortex and overlapped with those of many other objects. Thus repeating an object facilitates the recognition of the object while it impairs the recognition of other objects. These effects are called repetition priming and antipriming, respectively. Two experiments investigated a new phenomenon of repetition antipriming, in which a repeated object itself is antiprimed. The learning stage presented object pictures which were degraded at various levels. Participants determined how recognizable each object was. Then, the test stage presented the intact version of the object pictures and made participants to perform a categorization task. Both Experiment 1 and 2 found that the processing of the objects that had been recognized were facilitated (repetition priming) while the processing of the objects that had been perceptually ambiguous were impaired (repetition antipriming). These findings suggest that experiencing a perceptually ambiguous object might enhance the connection between feature-level representations and multiple object-level representations, which impairs the subsequent recognition of the repeated object.

• The KIPS Transactions:PartB
• /
• v.13B no.4 s.107
• /
• pp.429-438
• /
• 2006

Neural Networks(NNs) are applied for solving a wide variety of nonlinear problems in several areas, such as image processing, pattern recognition etc. Although NN can be simulated by using software, many potential NN applications required real-time processing. Thus they need to be implemented as hardware. The hardware implementation of multi-layer perceptrons(MLPs) in several kind of NNs usually uses a fixed-point arithmetic due to a simple logic operation and a shorter processing time compared to the floating-point arithmetic. However, the fixed-point arithmetic-based MLP has a drawback which is not able to apply the MLP software that use floating-point arithmetic. We propose a design method for MLPs which has the floating-point arithmetic-based fully-pipelining architecture. It has a processing speed that is proportional to the number of the hidden nodes. The number of input and output nodes of MLPs are generally constrained by given problems, but the number of hidden nodes can be optimized by user experiences. Thus our design method is using optimized number of hidden nodes in order to improve the processing speed, especially in field of a repeated processing such as image processing, pattern recognition, etc.

• The Korean Journal of Applied Statistics
• /
• v.31 no.1
• /
• pp.29-40
• /
• 2018

There are various reasons for missing values when collecting data. Missing values have some influence on the analysis and results; consequently, various methods of processing missing values have been studied to solve the problem. It is thought that the later point of view may be affected by the initial time point value in the repeated measurement data. However, in the existing method, there was no method for the imputation of missing values using this concept. Therefore, we proposed a new missing value imputation method in this study using clustering in initial time point of the repeated measurement data and the measure of property proposed by Kim and Kim (The Korean Communications in Statistics, 30, 463-473, 2017). We also applied the Monte Carlo simulations to compare the performance of the established method and suggested methods in repeated measurement data.

• Journal of KIISE
• /
• v.44 no.11
• /
• pp.1138-1148
• /
• 2017

Topic extraction is a technology that automatically extracts a set of topics from a set of documents, and this has been a major research topic in the area of natural language processing. Representative topic extraction methods include Latent Dirichlet Allocation (LDA) and word clustering-based methods. However, there are problems with these methods, such as repeated topics and mixed topics. The problem of repeated topics is one in which a specific topic is extracted as several topics, while the problem of mixed topic is one in which several topics are mixed in a single extracted topic. To solve these problems, this study proposes a method to extract topics using an LDA that is robust against the problem of repeated topic, going through the steps of separating and merging the topics using the similarity between words to correct the extracted topics. As a result of the experiment, the proposed method showed better performance than the conventional LDA method.

• Journal of KIISE:Software and Applications
• /
• v.34 no.7
• /
• pp.585-594
• /
• 2007

The back propagation algorithm took a long time to learn the input patterns and was difficult to train the additional or repeated learning patterns. So Aleksander proposed the binary neural network which could overcome the disadvantages of BP Network. But it had the limitation of repeated learning and was impossible to extract a generalized pattern. In this paper, we proposed a dynamic 3 dimensional Neuro System which was consisted of a learning network which was based on weightless neural network and a feedback module which could accumulate the characteristic. The proposed system was enable to train additional and repeated patterns. Also it could be produced a generalized pattern by putting a proper threshold into each learning-net's discriminator which was resulted from learning procedures. And then we reused the generalized pattern to elevate the recognition rate. In the last processing step to decide right category, we used maximum response detector. We experimented using the MNIST database of NIST and got 99.3% of right recognition rate for training data.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.22 no.1
• /
• pp.77-84
• /
• 2013

In this paper, we review two-dimensional measuring methods referring to target patterns. The patterns consist of two linearly-repeated patterns or is designed repeatedly in two-dimension. The repeated properties are reflectivity, refractivity, air-gapping distance, capacitance, magnetic reluctance, electrical resistance and sloping gradient, etc. However, the optical methods are generally used for high speed processing and density, and their encoding principles are treated here. In case of two-dimensional pattern, as there is not inherently error between single units encoding the pattern except for the metrology frame errors, the end-effector position of an object accompanying the pattern can be measured with respect of the global frame without via error. Therefore, it is regarded as a substitute for laser interferometer with severe environmental constraints and has been applied to the high-accurate planar actuator.

• The Korean Journal of Applied Statistics
• /
• v.30 no.6
• /
• pp.931-941
• /
• 2017

Typical Nonparametric methods for randomized block design with replications are two methods proposed by Mack (1981) and Mack and Skillings (1980). This method is likely to cause information loss because it uses the average of repeated observations instead of each repeated observation in the processing of each block. In order to compensate for this, we proposed a test method using linear placement statistics, which is a score function applied to the joint placement method proposed by Chung and Kim (2007). Monte Carlo simulation study is adapted to compare the power with previous methods.

• The KIPS Transactions:PartD
• /
• v.11D no.6
• /
• pp.1287-1294
• /
• 2004

Recently relationships between reliability measures and the coverage have been developed for evaluation of software reliability. Particularly the mean value function of the coverage-based software reliability growth model is important because of its key role in rep-resenting the software reliability growth. In this paper, we first review the problems of the existing mean value functions with respect to the assumptions on which they are based. Then a new mean value function is proposed. The new mean value function is developed for a general testing environment in which imperfect fault detection and repeated construct execution are allowed. Finally performance of the proposed model is empirically evaluated by applying it to a real data set.