• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.26-27
• /
• 2021

본 논문에서는 자막방송 제공을 위해 방송콘텐츠를 이해하는 방법으로 잔차 합성곱 순환신경망 기반 음향 사건 분류 기법을 제안한다. 제안된 기법은 잔차 합성곱 신경망과 순환 신경망을 연결한 구조를 갖는다. 신경망의 입력 특징으로는 멜-필터벵크 특징을 활용하고, 잔차 합성곱 신경망은 하나의 스템 블록과 5개의 잔차 합성곱 신경망으로 구성된다. 잔차 합성곱 신경망은 잔차 학습으로 구성된 합성곱 신경망과 기존의 합성곱 신경망 대비 특징맵의 표현 능력 향상을 위해 합성곱 블록 주의 모듈로 구성한다. 추출된 특징맵은 순환 신경망에 연결되고, 최종적으로 음향 사건 종류와 시간정보를 추출하는 완전연결층으로 연결되는 구조를 활용한다. 제안된 모델 훈련을 위해 라벨링되지 않는 데이터 활용이 가능한 평균 교사 모델을 기반으로 훈련하였다. 제안된 모델의 성능평가를 위해 DCASE 2020 챌린지 Task 4 데이터 셋을 활용하였으며, 성능 평가 결과 46.8%의 이벤트 단위의 F1-score를 얻을 수 있었다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.1
• /
• pp.1-8
• /
• 2018

Recently, the convolutional neural network shows high performance in many computer vision tasks. However, convolutional neural networks require enormous amount of operation, so it is difficult to adopt them in the embedded environments. To solve this problem, many studies are performed on the ASIC or FPGA implementation, where an efficient representation method is required. The fixed-point representation is adequate for the ASIC or FPGA implementation but causes a performance degradation. This paper proposes a separate optimization of representations for the convolutional layers and the batch normalization layers. With the proposed method, the required bit width for the convolutional layers is reduced from 16 bits to 10 bits for the ResNet-50 neural network. Since the computation amount of the convolutional layers occupies the most of the entire computation, the bit width reduction in the convolutional layers enables the efficient implementation of the convolutional neural networks.

• Transactions on Semiconductor Engineering
• /
• v.2 no.1
• /
• pp.1-8
• /
• 2024

When processing depthwise separable convolution, low utilization of processing elements (PEs) is one of the challenges of systolic array (SA). In this study, we propose a new SA architecture to maximize throughput in depthwise convolution. Moreover, the proposed SA performs subsequent pointwise convolution on the idle PEs during depthwise convolution computation to increase the utilization. After the computation, we utilize unused PEs to boost the remaining pointwise convolution. Consequently, the proposed 128x128 SA achieves a 4.05x and 1.75x speed improvement and reduces the energy consumption by 66.7 % and 25.4 %, respectively, compared to the basic SA and RiSA in MobileNetV3.

• Communications of Mathematical Education
• /
• v.37 no.2
• /
• pp.277-297
• /
• 2023

In mathematics, the concept of convolution is widely used. The convolution operation is required for understanding computer vision and deep learning in artificial intelligence. Therefore, it is vital for this concept to be explained in college mathematics education. In this paper, we present our new teaching and learning materials on convolution available for engineering mathematics. We provide the knowledge and applications on convolution with Python-based code, and introduce Convolutional Neural Network (CNN) used for image classification as an example. These materials can be utilized in class for the teaching of convolution and help students have a good understanding of the related knowledge in artificial intelligence.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.23 no.6
• /
• pp.38-44
• /
• 2019

A Convolution Neural Network(CNN) model was utilized to detect surface cracks in asphalt concrete pavements. The CNN used for this study consists of five layers with 3×3 convolution filter and 2×2 pooling kernel. Pavement surface crack images collected by automated road surveying equipment was used for the training and testing of the CNN. The performance of the CNN was evaluated using the accuracy, precision, recall, missing rate, and over rate of the surface crack detection. The CNN trained with the largest amount of data shows more than 96.6% of the accuracy, precision, and recall as well as less than 3.4% of the missing rate and the over rate.

• Communications of the Korean Mathematical Society
• /
• v.15 no.1
• /
• pp.1-27
• /
• 2000

위너공간과 추상 위너공간 위에서의 푸리에-파인만 변환, 푸리에-위너 변환 그리고 합성곱을 정의하고 여러가지 형태의 함수들에 대한 변환과 합성곱의 존재정리 및 여러가지 성질을 소개한다. 또한, 이 함수들의 파시발 관계와 프란셰렐 관계에 대해서도 알아본다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.11a
• /
• pp.34-35
• /
• 2017

본 논문에서는 영상으로부터 생성된 깊이맵을 합성곱 신경망(CNN)으로 재생성하는 방법을 제안한다. 합성곱 신경망은 영상인식, 영상분류에 좋은 성능을 보여주는데, 이 기술을 깊이맵 생성에 활용하여 기 제작된 깊이맵 생성 기법을 간단한 합성곱 신경망으로 구현하고자 한다. 성능 실험에서는 10개의 비디오 세트에 제안 방법을 적용한 결과, 만족스러운 결과를 얻었다.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.4
• /
• pp.29-37
• /
• 2021

In this paper, we propose a deep learning-based retinal vessel segmentation model for handling multi-scale information of fundus images. we integrate the selective kernel convolution into U-Net-based convolutional neural network. The proposed model extracts and segment features information with various shapes and sizes of retinal blood vessels, which is important information for diagnosing eye-related diseases from fundus images. The proposed model consists of standard convolutions and selective kernel convolutions. While the standard convolutional layer extracts information through the same size kernel size, The selective kernel convolution extracts information from branches with various kernel sizes and combines them by adaptively adjusting them through split-attention. To evaluate the performance of the proposed model, we used the DRIVE and CHASE DB1 datasets and the proposed model showed F1 score of 82.91% and 81.71% on both datasets respectively, confirming that the proposed model is effective in segmenting retinal blood vessels.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.397-399
• /
• 2022

Digit recognition is one of the applications of deep learning, which appears in many fields. CNN network enables us to recognize handwritten digits. Also, It can process various types of data. As we stack more layers in CNN network, we expect more performance improvements. In this paper, we added a convolution layer. as a result, we achieved an accuracy improvement from 76.96% to 98.87%, which is a nearly 21.81% increase.

• Journal of Korean Society of Disaster and Security
• /
• v.8 no.1
• /
• pp.15-19
• /
• 2015

In rainfall-runoff models employed in hydrological applications, runoff amount is estimated through temporal delay of effective precipitation based on a linear system. Its amount is resulted from the linearized ratio by analyzing the convolution multiplier. Furthermore, in case of kernel density estimate (KDE) used in probabilistic analysis, the definition of the kernel comes from the convolution multiplier. Individual data values are smoothed through the kernel to derive KDE. In the current study, the roles of the convolution multiplier for KDE and rainfall-runoff models were revisited and their similarity and dissimilarity were investigated to discover the mathematical applicability of the convolution multiplier.