• 전기학회논문지
• /
• 제67권12호
• /
• pp.1665-1670
• /
• 2018

A convolutional neural network (CNN), which is one of the deep learning models, has been very successful in a variety of computer vision tasks. Filters of a CNN are automatically generated, however, they can be further optimized since there exist the possibility of existing redundant and less important features. Therefore, the aim of this paper is a filter reduction to accelerate and compress CNN models. Evolutionary algorithms is adopted to remove the unnecessary filters in order to minimize the parameters of CNN networks while maintaining a good performance of classification. We demonstrate the proposed filter reduction methods performing experiments on CIFAR10 data based on the classification performance. The comparison for three approaches is analysed and the outlook for the potential next steps is suggested.

• Journal of Multimedia Information System
• /
• 제6권4호
• /
• pp.165-172
• /
• 2019

Machine-learning techniques have been actively employed to information security in recent years. Traditional rule-based security solutions are vulnerable to advanced attacks due to unpredictable behaviors and unknown vulnerabilities. By employing ML techniques, we are able to develop intrusion detection systems (IDS) based on anomaly detection instead of misuse detection. Moreover, threshold issues in anomaly detection can also be resolved through machine-learning. There are very few datasets for network intrusion detection compared to datasets for malicious code. KDD CUP 99 (KDD) is the most widely used dataset for the evaluation of IDS. Numerous studies on ML-based IDS have been using KDD or the upgraded versions of KDD. In this work, we develop an IDS model using CSE-CIC-IDS 2018, a dataset containing the most up-to-date common network attacks. We employ deep-learning techniques and develop a convolutional neural network (CNN) model for CSE-CIC-IDS 2018. We then evaluate its performance comparing with a recurrent neural network (RNN) model. Our experimental results show that the performance of our CNN model is higher than that of the RNN model when applied to CSE-CIC-IDS 2018 dataset. Furthermore, we suggest a way of improving the performance of our model.

• 전기전자학회논문지
• /
• 제20권4호
• /
• pp.351-360
• /
• 2016

최근 딥 러닝을 중심으로 빠르게 발전하고 있는 기계학습 분류 알고리즘은 기존의 방법들보다 뛰어난 성능으로 인하여 주목받고 있다. 딥 러닝 중에서도 Convolutional Neural Network(CNN)는 영상처리에 뛰어나 첨단 운전자 보조 시스템(Advanced Driver Assistance System : ADAS)에서 많이 사용되고 있는 추세이다. 하지만 차량용 임베디드 환경에서 CNN을 소프트웨어로 동작시켰을 때는 각 Layer마다 연산이 반복되는 알고리즘의 특성으로 인해 수행시간이 길어져 실시간 처리가 어렵다. 본 논문에서는 임베디드 환경에서 CNN의 실시간 처리를 위하여 Convolution 연산 및 기타 연산들을 병렬로 처리하여 CNN의 속도를 향상시키는 하드웨어 구조를 제안한다. 제안하는 하드웨어의 성능을 검증하기 위하여 Xilinx ZC706 FPGA 보드를 이용하였다. 입력 영상은 $36{\times}36$ 크기이며, 동작주파수 100MHz에서 하드웨어 수행시간은 약 2.812ms로 실시간 처리가 가능함을 확인했다.

• 한국멀티미디어학회논문지
• /
• 제25권11호
• /
• pp.1643-1652
• /
• 2022

Most of the existing machine reading research has used Recurrent Neural Network (RNN) and Convolutional Neural Network (CNN) algorithms as networks. Among them, RNN was slow in training, and Question Answering Network (QANet) was announced to improve training speed. QANet is a model composed of CNN and self-attention. CNN extracts semantic and syntactic information well from the local corpus, but there is a limit to extracting the corresponding information from the global corpus. Graph Convolutional Networks (GCN) extracts semantic and syntactic information relatively well from the global corpus. In this paper, to take advantage of this strength of GCN, we propose I-QANet, which changed the CNN of QANet to GCN. The proposed model performed 1.2 times faster than the baseline in the Stanford Question Answering Dataset (SQuAD) dataset and showed 0.2% higher performance in Exact Match (EM) and 0.7% higher in F1. Furthermore, in the Korean Question Answering Dataset (KorQuAD) dataset consisting only of Korean, the learning time was 1.1 times faster than the baseline, and the EM and F1 performance were also 0.9% and 0.7% higher, respectively.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권2호
• /
• pp.693-709
• /
• 2018

In this paper, we propose a compact Convolutional Neural Network (CNN)-based tracker in conjunction with a particle filter architecture, in which the CNN model operates as an accurate candidates estimator, while the particle filter predicts the target motion dynamics, lowering the overall number of calculations and refines the resulting target bounding box. Experiments were conducted on the Online Object Tracking Benchmark (OTB) [34] dataset and comparison analysis in respect to other state-of-art has been performed based on accuracy and precision, indicating that the proposed algorithm outperforms all state-of-the-art trackers included in the OTB dataset, specifically, TLD [16], MIL [1], SCM [36] and ASLA [15]. Also, a comprehensive speed performance analysis showed average frames per second (FPS) among the top-10 trackers from the OTB dataset [34].

• ETRI Journal
• /
• 제43권4호
• /
• pp.684-693
• /
• 2021

This paper presents a low-cost two-stage approximate multiplier for bfloat16 (brain floating-point) data processing. For cost-efficient approximate multiplication, the first stage implements Mitchell's algorithm that performs the approximate multiplication using only two adders. The second stage adopts the exact multiplication to compensate for the error from the first stage by multiplying error terms and adding its truncated result to the final output. In our design, the low-cost multiplications in both stages can reduce hardware costs significantly and provide low relative errors by compensating for the error from the first stage. We apply our approximate multiplier to the convolutional neural network (CNN) inferences, which shows small accuracy drops with well-known pre-trained models for the ImageNet database. Therefore, our design allows low-cost CNN inference systems with high test accuracy.

• 한국컴퓨터정보학회논문지
• /
• 제24권1호
• /
• pp.33-39
• /
• 2019

In this paper, we propose deep convolutional neural network(CNN) with bottleneck structure which improves the performance of earthquake classification. In order to address all possible forms of earthquakes including micro-earthquakes and artificial-earthquakes as well as large earthquakes, we need a representation and classifier that can effectively discriminate seismic waveforms in adverse conditions. In particular, to robustly classify seismic waveforms even in low snr, a deep CNN with 1x1 convolution bottleneck structure is proposed in raw seismic waveforms. The representative experimental results show that the proposed method is effective for noisy seismic waveforms and outperforms the previous state-of-the art methods on domestic earthquake database.

• Geomechanics and Engineering
• /
• 제36권1호
• /
• pp.1-8
• /
• 2024

We propose a new method for detecting voids behind tunnel concrete linings using the impact-echo method that is based on continuous wavelet transform (CWT) and a convolutional neural network (CNN). We first collect experimental data using the impact-echo method and then convert them into time-frequency images via CWT. We provide a CNN model trained using the converted images and experimentally confirm that our proposed model is robust. Moreover, it exhibits outstanding performance in detecting backfill voids and their status.

• 스마트미디어저널
• /
• 제9권1호
• /
• pp.23-29
• /
• 2020

This paper presents an approach for dynamic hand gesture recognition by using algorithm based on 3D Convolutional Neural Network (3D_CNN), which is later extended to 3D Residual Networks (3D_ResNet), and the neural network based key frame selection. Typically, 3D deep neural network is used to classify gestures from the input of image frames, randomly sampled from a video data. In this work, to improve the classification performance, we employ key frames which represent the overall video, as the input of the classification network. The key frames are extracted by SegNet instead of conventional clustering algorithms for video summarization (VSUMM) which require heavy computation. By using a deep neural network, key frame selection can be performed in a real-time system. Experiments are conducted using 3D convolutional kernels such as 3D_CNN, Inflated 3D_CNN (I3D) and 3D_ResNet for gesture classification. Our algorithm achieved up to 97.8% of classification accuracy on the Cambridge gesture dataset. The experimental results show that the proposed approach is efficient and outperforms existing methods.

• 정보보호학회논문지
• /
• 제30권6호
• /
• pp.1271-1278
• /
• 2020

딥러닝 기반 프로파일링 부채널 분석은 신경망을 이용해 부채널 정보와 중간값의 관계를 파악하는 공격 방법이다. 신경망은 신호의 각 시점을 별도의 차원으로 해석하므로 차원별 가중치를 갖는 신경망은 지터가 있는 데이터의 분포를 학습하기 어렵다. 본 논문에서는 CNN(Convolutional Neural Network)의 완전연결 층을 GAP(Global Average Pooling)로 대체하면 태생적으로 지터에 강건한 신경망을 구성할 수 있음을 보인다. 이를 입증하기 위해 ChipWhisperer-Lite 전력 수집 보드에서 수집한 파형에 대해 실험한 결과 검증 데이터 집합에 대한 완전연결 층을 사용하는 CNN의 정확도는 최대 1.4%에 불과했으나, GAP를 사용하는 CNN의 정확도는 최대 41.7%로 매우 높게 나타났다.