• Journal of Korea Multimedia Society
• /
• v.22 no.6
• /
• pp.665-675
• /
• 2019

In this paper, we compare and analyze the classification performance of deep learning algorithm Convolutional Neural Network(CNN) ac cording to ensemble generation and combining techniques. We used several CNN models(VGG16, VGG19, DenseNet121, DenseNet169, DenseNet201, ResNet18, ResNet34, ResNet50, ResNet101, ResNet152, GoogLeNet) to create 10 ensemble generation combinations and applied 6 combine techniques(average, weighted average, maximum, minimum, median, product) to the optimal combination. Experimental results, DenseNet169-VGG16-GoogLeNet combination in ensemble generation, and the product rule in ensemble combination showed the best performance. Based on this, it was concluded that ensemble in different models of high benchmarking scores is another way to get good results.

• Journal of the Semiconductor & Display Technology
• /
• v.20 no.3
• /
• pp.125-130
• /
• 2021

Scaffold is used to produce bio sensor. Scaffold is required high dimensional accuracy. 3D printer is used to manufacture scaffold. 3D printer can't detect defect during printing. Defect detection is very important in scaffold printing. Real-time defect detection is very necessary on industry. In this paper, we proposed the method for real-time scaffold defect detection. Real-time defect detection model is produced using CNN(Convolution Neural Network) algorithm. Performance of the proposed model has been verified through evaluation. Real-time defect detection system are manufactured on hardware. Experiments were conducted to detect scaffold defects in real-time. As result of verification, the defect detection system detected scaffold defect well in real-time.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2016.10a
• /
• pp.405-408
• /
• 2016

Games like Go, Chess, Janggi have helped to brain development of the people. These games are developed by computer program. And many algorithms have been developed to allow myself to play. The person winning chess program was developed in the 1990s. But game of go is too large number of cases. So it was considered impossible to win professional go player. However, with the use of MCTS(Monte Carlo Tree Search) and CNN(Convolutional Neural Network), the performance of the go algorithm is greatly improved. In this paper, using CNN and MCTS were proceeding development of go algorithm. Using the manual of go learning CNN look for the best position, MCTS calculates the win probability in the game to proceed with simulation. In addition, extract pattern information of go using existing manual of go, plans to improve speed and performance by using it. This method is showed a better performance than general go algorithm. Also if it is receiving sufficient computing power, it seems to be even more improved performance.

• ETRI Journal
• /
• v.27 no.3
• /
• pp.294-303
• /
• 2005

In this paper, a supervised algorithm for the evaluation of geophysical sites using a multi-level cellular neural network (ML-CNN) is introduced, developed, and applied to real data. ML-CNN is a stochastic image processing technique based on template optimization using neighborhood relationships of the pixels. The separation/enhancement and border detection performance of the proposed method is evaluated by various interesting real applications. A genetic algorithm is used in the optimization of CNN templates. The first application is concerned with the separation of potential field data of the Dumluca chromite region, which is one of the rich reserves of Turkey; in this context, the classical approach to the gravity anomaly separation method is one of the main problems in geophysics. The other application is the border detection of archeological ruins of the Hittite Empire in Turkey. The Hittite civilization sites located at the Sivas-Altinyayla region of Turkey are among the most important archeological sites in history, one reason among others being that written documentation was first produced by this civilization.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.10 no.4
• /
• pp.288-294
• /
• 2017

Recently, deep learning using convolutional neural network (CNN) has been extensively studied in image recognition. Convolution consists of addition and multiplication. Multiplication is computationally expensive in hardware implementation, relative to addition. It is also important factor limiting a chip design in an embedded deep learning system. In this paper, I propose a parallel-addition processing algorithm that converts grayscale images to the superposition of binary images and performs convolution only with addition. It is confirmed that the convolution can be performed by a parallel-addition method capable of reducing the processing time in experiment for verifying the availability of proposed algorithm.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.18 no.12
• /
• pp.38-44
• /
• 2019

Various automation studies have been conducted to detect defective products based on product images. In the case of machine vision-based studies, size and color error are detected through a preprocessing process. A situation may arise in which the main features are removed during the preprocessing process, thereby decreasing the accuracy. In addition, complex systems are required to detect various kinds of defects. In this study, we designed and developed a system to detect errors by analyzing various conditions of defective products. We designed the deep learning algorithm to detect the defective features from the product images during the automation process using a convolution neural network (CNN) and verified the performance by applying the algorithm to the checker-switch failure detection system. It was confirmed that all seven error characteristics were detected accurately, and it is expected that it will show excellent performance when applied to automation systems for error detection.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.6
• /
• pp.704-709
• /
• 2003

The digital image can be distorted by a noise for a transmission or other elements of system. It happen to be vague of a boundary side in the division of an image object, especially, boundary side of an input image is very important because it can be determined to the division and detection element in pattern recognition. Therefore it is proposed an edge detection method of optimal to divide and detect exactly a boundary part. In this paper, it detected the optimal edge with applying this image to Meyer wavelet-CNN algorithm, after it does level up a boundary side of an image by using the adaptive morphology as the threshold of an input image. It confirmed that the proposed algorithm is more superior to the conventional methods and the conventional Sobel method which is an image edge detection algorithm. Especially, it is confirmed by simulation that the proposed algorithm can be got the better result edge at the place of closing to each edges and having smoothly curved line.

• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.217-226
• /
• 2019

In this paper, we propose a filtering method that can improve the results of inverse tone-mapping using guided image filter. Inverse tone-mapping techniques have been proposed that convert LDR images to HDR. Recently, many algorithms have been studied to convert single LDR images into HDR images using CNN. Among them, there exists an algorithm for restoring pixel information using CNN which learned to restore saturated region. The algorithm does not suppress the noise in the non-saturation region and cannot restore the detail in the saturated region. The proposed algorithm suppresses the noise in the non-saturated region and restores the detail of the saturated region using a WGIF in the input image, and then applies it to the CNN to improve the quality of the final image. The proposed algorithm shows a higher quantitative image quality index than the existing algorithms when the HDR quantitative image quality index was measured.

• Journal of IKEEE
• /
• v.25 no.1
• /
• pp.180-192
• /
• 2021

One widely used algorithm for image recognition and pattern detection is the convolution neural network (CNN). To efficiently handle convolution operations, which account for the majority of computations in the CNN, we use hardware accelerators to improve the performance of CNN applications. In using these hardware accelerators, the CNN fetches data from the off-chip DRAM, as the massive computational volume of data makes it difficult to derive performance improvements only from memory inside the hardware accelerator. In other words, data communication between off-chip DRAM and memory inside the accelerator has a significant impact on the performance of CNN applications. In this paper, a simulator for the CNN is developed to analyze the main memory or DRAM access with respect to the size of the on-chip memory or global buffer inside the CNN accelerator. For AlexNet, one of the CNN architectures, when simulated with increasing the size of the global buffer, we found that the global buffer of size larger than 100kB has 0.8x as low a DRAM access count as the global buffer of size smaller than 100kB.

• Journal of Broadcast Engineering
• /
• v.23 no.4
• /
• pp.484-494
• /
• 2018

The uniform motion blur removing algorithm of Cho et al. has the problem that the edge region of the image cannot be restored clearly. We propose the effective algorithm to overcome this problem by using shock filter that reconstructs a blurred step signal into a sharp edge, and convolutional neural network (CNN) that learns by extracting features from the image. Then uniform motion blur kernel is estimated from the latent sharp image to remove blur in the image. The proposed algorithm improved the disadvantages of the conventional algorithm by reconstructing the latent sharp image using shock filter and CNN. Through the experimental results, it was confirmed that the proposed algorithm shows excellent reconstruction performance in objective and subjective image quality than the conventional algorithm.