• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.40 no.1
• /
• pp.1-9
• /
• 2003

The CNN-UM algorithm which performs the analog parallel subtraction of images has been developed and its application study to the moving target detection has been done. The CNN-UM is the state of the art computation architecture with high computational potential of analog parallel processing. It is one of the strong candidates for the next generation of computing system which fulfills requirement of the real-time image processing. One weakness of the CNN-UM is that its analog parallel processing function is not fully utilized for the inter frame processing. If two subsequent image frames are superimposed with opposite signs on identical capacitors for short time period, the analog subtraction between them is achieved. The Principle of such temporal inter-frame processing algorithm has been described and its mathematical analysis has been done. Practical usefulness of the proposed algorithm has also been verified through the application for moving target detection.

• ETRI Journal
• /
• v.45 no.4
• /
• pp.666-677
• /
• 2023

In the last few years, convolutional neural networks (CNNs) have demonstrated good performance while solving various computer vision problems. However, since CNNs exhibit high computational complexity, signal processing is performed on the server side. To reduce the computational complexity of CNNs for edge computing, a lightweight algorithm, such as a MobileNet, is proposed. Although MobileNet is lighter than other CNN models, it commonly achieves lower classification accuracy. Hence, to find a balance between complexity and accuracy, additional hyperparameters for adjusting the size of the model have recently been proposed. However, significantly increasing the number of parameters makes models dense and unsuitable for devices with limited computational resources. In this study, we propose a novel MobileNet architecture, in which the number of parameters is adaptively increased according to the importance of feature maps. We show that our proposed network achieves better classification accuracy with fewer parameters than the conventional MobileNet.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.2
• /
• pp.113-118
• /
• 2023

The WiFi fingerprint method for location estimation within the home has the advantage of using the existing infrastructure and estimating absolute coordinates, so many studies are being conducted. Existing studies have mainly focused on the study of localization algorithms, but the improvement of accuracy has reached its limits. However, since a wireless LAN receiver such as a smartphone cannot measure signals smaller than the reception sensitivity of radio signals, the position estimation error varies depending on the method of processing these values. In this paper, we proposed a method to increase the location estimation accuracy by pre-processing the received signal data of the measured wireless LAN router in various ways and applying it to the existing algorithm, and greatly improved accuracy was obtained. In addition, the preprocessed data was applied to the KNN method and the CNN method and the performance was compared.

• Journal of Semiconductor Engineering
• /
• v.2 no.2
• /
• pp.130-135
• /
• 2021

Recently, deep neural networks (DNNs) are actively used for action control so that an autonomous system, such as the robot, can perform human-like behaviors and operations. Unlike recognition tasks, the real-time operation is essential in action control, and it is too slow to use remote learning on a server communicating through a network. New learning techniques, such as reinforcement learning (RL), are needed to determine and select the correct robot behavior locally. In this paper, we propose an energy-efficient DNN processor with a LUT-based processing engine and near-zero skipper. A CNN-based facial emotion recognition and an RNN-based emotional dialogue generation model is integrated for natural HRI system and tested with the proposed processor. It supports 1b to 16b variable weight bit precision with and 57.6% and 28.5% lower energy consumption than conventional MAC arithmetic units for 1b and 16b weight precision. Also, the near-zero skipper reduces 36% of MAC operation and consumes 28% lower energy consumption for facial emotion recognition tasks. Implemented in 65nm CMOS process, the proposed processor occupies 1784×1784 um2 areas and dissipates 0.28 mW and 34.4 mW at 1fps and 30fps facial emotion recognition tasks.