• Title/Summary/Keyword: 경량화 딥러닝 모델

Search Result 56, Processing Time 0.04 seconds

RoutingConvNet: A Light-weight Speech Emotion Recognition Model Based on Bidirectional MFCC (RoutingConvNet: 양방향 MFCC 기반 경량 음성감정인식 모델)

  • Hyun Taek Lim;Soo Hyung Kim;Guee Sang Lee;Hyung Jeong Yang
    • Smart Media Journal
    • /
    • v.12 no.5
    • /
    • pp.28-35
    • /
    • 2023
  • In this study, we propose a new light-weight model RoutingConvNet with fewer parameters to improve the applicability and practicality of speech emotion recognition. To reduce the number of learnable parameters, the proposed model connects bidirectional MFCCs on a channel-by-channel basis to learn long-term emotion dependence and extract contextual features. A light-weight deep CNN is constructed for low-level feature extraction, and self-attention is used to obtain information about channel and spatial signals in speech signals. In addition, we apply dynamic routing to improve the accuracy and construct a model that is robust to feature variations. The proposed model shows parameter reduction and accuracy improvement in the overall experiments of speech emotion datasets (EMO-DB, RAVDESS, and IEMOCAP), achieving 87.86%, 83.44%, and 66.06% accuracy respectively with about 156,000 parameters. In this study, we proposed a metric to calculate the trade-off between the number of parameters and accuracy for performance evaluation against light-weight.

Lightweight Deep Learning Model for Real-Time 3D Object Detection in Point Clouds (실시간 3차원 객체 검출을 위한 포인트 클라우드 기반 딥러닝 모델 경량화)

  • Kim, Gyu-Min;Baek, Joong-Hwan;Kim, Hee Yeong
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.26 no.9
    • /
    • pp.1330-1339
    • /
    • 2022
  • 3D object detection generally aims to detect relatively large data such as automobiles, buses, persons, furniture, etc, so it is vulnerable to small object detection. In addition, in an environment with limited resources such as embedded devices, it is difficult to apply the model because of the huge amount of computation. In this paper, the accuracy of small object detection was improved by focusing on local features using only one layer, and the inference speed was improved through the proposed knowledge distillation method from large pre-trained network to small network and adaptive quantization method according to the parameter size. The proposed model was evaluated using SUN RGB-D Val and self-made apple tree data set. Finally, it achieved the accuracy performance of 62.04% at mAP@0.25 and 47.1% at mAP@0.5, and the inference speed was 120.5 scenes per sec, showing a fast real-time processing speed.

Addressing Inter-floor Noise Issues in Apartment Buildings using On-Sensor AI Embedded with TinyML on Ultra-Low-Power Systems

  • Jae-Won Kwak;In-Yeop Choi
    • Journal of the Korea Society of Computer and Information
    • /
    • v.29 no.3
    • /
    • pp.75-81
    • /
    • 2024
  • In this paper, we proposes a method for real-time processing of inter-floor noise problems by embedding TinyML, which includes a deep learning model, into ultra-low-power systems. The reason this method is feasible is because of lightweight deep learning model technology, which allows even systems with small computing resources to perform inference autonomously. The conventional method proposed to solve inter-floor noise problems was to send data collected from sensors to a server for analysis and processing. However, this centralized processing method has issues with high costs, complexity, and difficulty in real-time processing. In this paper, we address these limitations by employing On-Sensor AI using TinyML. The method presented in this paper is simple to install, cost-effective, and capable of processing problems in real-time.

An Efficient Data Collection Method for Deep Learning-based Wireless Signal Identification in Unlicensed Spectrum (딥 러닝 기반의 이기종 무선 신호 구분을 위한 데이터 수집 효율화 기법)

  • Choi, Jaehyuk
    • Journal of IKEEE
    • /
    • v.26 no.1
    • /
    • pp.62-66
    • /
    • 2022
  • Recently, there have been many research efforts based on data-based deep learning technologies to deal with the interference problem between heterogeneous wireless communication devices in unlicensed frequency bands. However, existing approaches are commonly based on the use of complex neural network models, which require high computational power, limiting their efficiency in resource-constrained network interfaces and Internet of Things (IoT) devices. In this study, we address the problem of classifying heterogeneous wireless technologies including Wi-Fi and ZigBee in unlicensed spectrum bands. We focus on a data-driven approach that employs a supervised-learning method that uses received signal strength indicator (RSSI) data to train Deep Convolutional Neural Networks (CNNs). We propose a simple measurement methodology for collecting RSSI training data which preserves temporal and spectral properties of the target signal. Real experimental results using an open-source 2.4 GHz wireless development platform Ubertooth show that the proposed sampling method maintains the same accuracy with only a 10% level of sampling data for the same neural network architecture.

REALM for Open-domain Question Answering of Korean (REALM을 이용한 한국어 오픈도메인 질의 응답)

  • Kan, Dong-Chan;Na, Seung-Hoon;Choi, Yun-Su;Lee, Hye-Woo;Chang, Du-Seong
    • Annual Conference on Human and Language Technology
    • /
    • 2020.10a
    • /
    • pp.192-196
    • /
    • 2020
  • 최근 딥러닝 기술의 발전에 힘입어 오픈 도메인 QA 시스템의 발전은 가속화되고 있다. 특히 IR 시스템(Information Retrieval)과 추출 기반의 기계 독해 모델을 결합한 접근 방식(IRQA)의 경우, 문서와 질문 각각을 연속 벡터로 인코딩하는 IR 시스템(Dense Retrieval)의 연구가 진행되면서 검색 성능이 전통적인 키워드 기반 IR 시스템에 비해 큰 폭으로 상승하였고, 이를 기반으로 오픈 도메인 질의응답의 성능 또한 개선 되었다. 본 논문에서는 경량화 된 BERT 모델을 기반으로 하여 Dense Retrieval 모델 ORQA와 REALM을 사전 학습하고, 한국어 오픈 도메인 QA에서 QA 성능과 검색 성능을 도출한다. 실험 결과, 키워드 기반 IR 시스템 BM25를 기반으로 했던 이전 IRQA 실험결과와 비교하여 더 적은 문서로 더 나은 QA 성능을 보였으며, 검색 결과의 경우, BM25의 성능을 뛰어넘는 결과를 보였다.

  • PDF

An Evaluation of Inference Acceleration for Drone-based Real-time Object Detection (드론 기반 실시간 객체 식별을 위한 추론 가속화 평가)

  • Kwon, Seung-Sang;Moon, Yong-Hyuk
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2022.11a
    • /
    • pp.408-410
    • /
    • 2022
  • 최근 데이터 획득 위치에 가장 근접하고, 저 수준의 계산력을 제공하는 엣지 기기를 중심으로 직접 딥러닝 추론을 수행하고자 하는 요구가 증가하고 있다. 본 논문에서는 드론에서 촬영한 교통 영상 데이터를 기반으로, 다수의 차량 종류 및 보행자를 식별하는 모델을 Jetson Nano 에 탑재하여 기본 성능을 측정한다. 더불어, 자원제약형 기기 환경에서 TensorRT 와 Deepstream 을 활용하여 객체 식별 모델의 연산 경량화 및 추론 가속화 성능을 극대화하기 위한 구현 및 실험을 수행하여 Anchor-based 및 Anchor-free 객체 식별 모델의 정확도와 실시간 대응력을 평가하고 논의한다.

Acceleration of CNN Model Using Neural Network Compression and its Performance Evaluation on Embedded Boards (임베디드 보드에서의 인공신경망 압축을 이용한 CNN 모델의 가속 및 성능 검증)

  • Moon, Hyeon-Cheol;Lee, Ho-Young;Kim, Jae-Gon
    • Proceedings of the Korean Society of Broadcast Engineers Conference
    • /
    • 2019.11a
    • /
    • pp.44-45
    • /
    • 2019
  • 최근 CNN 등 인공신경망은 최근 이미지 분류, 객체 인식, 자연어 처리 등 다양한 분야에서 뛰어난 성능을 보이고 있다. 그러나, 대부분의 분야에서 보다 더 높은 성능을 얻기 위해 사용한 인공신경망 모델들은 파라미터 수 및 연산량 등이 방대하여, 모바일 및 IoT 디바이스 같은 연산량이나 메모리가 제한된 환경에서 추론하기에는 제한적이다. 따라서 연산량 및 모델 파라미터 수를 압축하기 위한 딥러닝 경량화 알고리즘이 연구되고 있다. 본 논문에서는 임베디트 보드에서의 압축된 CNN 모델의 성능을 검증한다. 인공지능 지원 맞춤형 칩인 QCS605 를 내장한 임베디드 보드에서 카메라로 입력한 영상에 대해서 원 CNN 모델과 압축된 CNN 모델의 분류 성능과 동작속도 비교 분석한다. 본 논문의 실험에서는 CNN 모델로 MobileNetV2, VGG16 을 사용했으며, 주어진 모델에서 가지치기(pruning) 기법, 양자화, 행렬 분해 등의 인공신경망 압축 기술을 적용하였을 때 원래의 모델 대비 추론 시간 및 분류의 정확도 성능을 분석하고 인공신경망 압축 기술의 유용성을 확인하였다.

  • PDF

Detection Algorithm of Road Surface Damage Using Adversarial Learning (적대적 학습을 이용한 도로 노면 파손 탐지 알고리즘)

  • Shim, Seungbo
    • The Journal of The Korea Institute of Intelligent Transport Systems
    • /
    • v.20 no.4
    • /
    • pp.95-105
    • /
    • 2021
  • Road surface damage detection is essential for a comfortable driving environment and the prevention of safety accidents. Road management institutes are using automated technology-based inspection equipment and systems. As one of these automation technologies, a sensor to detect road surface damage plays an important role. For this purpose, several studies on sensors using deep learning have been conducted in recent years. Road images and label images are needed to develop such deep learning algorithms. On the other hand, considerable time and labor will be needed to secure label images. In this paper, the adversarial learning method, one of the semi-supervised learning techniques, was proposed to solve this problem. For its implementation, a lightweight deep neural network model was trained using 5,327 road images and 1,327 label images. After experimenting with 400 road images, a model with a mean intersection over a union of 80.54% and an F1 score of 77.85% was developed. Through this, a technology that can improve recognition performance by adding only road images was developed to learning without label images and is expected to be used as a technology for road surface management in the future.

Vehicle License Plate Recognition System using SSD-Mobilenet and ResNet for Mobile Device (SSD-Mobilenet과 ResNet을 이용한 모바일 기기용 자동차 번호판 인식시스템)

  • Kim, Woonki;Dehghan, Fatemeh;Cho, Seongwon
    • Smart Media Journal
    • /
    • v.9 no.2
    • /
    • pp.92-98
    • /
    • 2020
  • This paper proposes a vehicle license plate recognition system using light weight deep learning models without high-end server. The proposed license plate recognition system consists of 3 steps: [license plate detection]-[character area segmentation]-[character recognition]. SSD-Mobilenet was used for license plate detection, ResNet with localization was used for character area segmentation, ResNet was used for character recognition. Experiemnts using Samsung Galaxy S7 and LG Q9, accuracy showed 85.3% accuracy and around 1.1 second running time.

A Study on Lightweight Model with Attention Process for Efficient Object Detection (효율적인 객체 검출을 위해 Attention Process를 적용한 경량화 모델에 대한 연구)

  • Park, Chan-Soo;Lee, Sang-Hun;Han, Hyun-Ho
    • Journal of Digital Convergence
    • /
    • v.19 no.5
    • /
    • pp.307-313
    • /
    • 2021
  • In this paper, a lightweight network with fewer parameters compared to the existing object detection method is proposed. In the case of the currently used detection model, the network complexity has been greatly increased to improve accuracy. Therefore, the proposed network uses EfficientNet as a feature extraction network, and the subsequent layers are formed in a pyramid structure to utilize low-level detailed features and high-level semantic features. An attention process was applied between pyramid structures to suppress unnecessary noise for prediction. All computational processes of the network are replaced by depth-wise and point-wise convolutions to minimize the amount of computation. The proposed network was trained and evaluated using the PASCAL VOC dataset. The features fused through the experiment showed robust properties for various objects through a refinement process. Compared with the CNN-based detection model, detection accuracy is improved with a small amount of computation. It is considered necessary to adjust the anchor ratio according to the size of the object as a future study.