• 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.

• Korean Journal of Artificial Intelligence
• /
• v.11 no.4
• /
• pp.1-8
• /
• 2023

In this paper, we propose a black ice detection platform framework using Convolutional Neural Networks (CNNs). To overcome black ice problem, we introduce a real-time based early warning platform using CNN-based architecture, and furthermore, in order to enhance the accuracy of black ice detection, we apply a multi-scale dilation convolution feature fusion (MsDC-FF) technique. Then, we establish a specialized experimental platform by using a comprehensive dataset of thermal road black ice images for a training and evaluation purpose. Experimental results of a real-time black ice detection platform show the better performance of our proposed network model compared to conventional image segmentation models. Our proposed platform have achieved real-time segmentation of road black ice areas by deploying a road black ice area segmentation network on the edge device Jetson Nano devices. This approach in parallel using multi-scale dilated convolutions with different dilation rates had faster segmentation speeds due to its smaller model parameters. The proposed MsCD-FF Net(2) model had the fastest segmentation speed at 5.53 frame per second (FPS). Thereby encouraging safe driving for motorists and providing decision support for road surface management in the road traffic monitoring department.

• Journal of Information Processing Systems
• /
• v.14 no.6
• /
• pp.1494-1507
• /
• 2018

Intelligent human identification using face information has been the research hotspot ranging from Internet of Things (IoT) application, intelligent self-service bank, intelligent surveillance to public safety and intelligent access control. Since 2D face images are usually captured from a long distance in an unconstrained environment, to fully exploit this advantage and make human recognition appropriate for wider intelligent applications with higher security and convenience, the key difficulties here include gray scale change caused by illumination variance, occlusion caused by glasses, hair or scarf, self-occlusion and deformation caused by pose or expression variation. To conquer these, many solutions have been proposed. However, most of them only improve recognition performance under one influence factor, which still cannot meet the real face recognition scenario. In this paper we propose a multi-scale parallel convolutional neural network architecture to extract deep robust facial features with high discriminative ability. Abundant experiments are conducted on CMU-PIE, extended FERET and AR database. And the experiment results show that the proposed algorithm exhibits excellent discriminative ability compared with other existing algorithms.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.4
• /
• pp.234-240
• /
• 2022

This paper proposes a customized AI exercise recommendation service for balancing the relative amount of exercise according to the working environment by each occupation. WISDM database is collected by using acceleration and gyro sensors, and is a dataset that classifies physical activities into 18 categories. Our system recommends a adaptive exercise using the analyzed activity type after classifying 18 physical activities into 3 physical activities types such as whole body, upper body and lower body. 1 Dimensional convolutional neural network is used for classifying a physical activity in this paper. Proposed model is composed of a convolution blocks in which 1D convolution layers with a various sized kernel are connected in parallel. Convolution blocks can extract a detailed local features of input pattern effectively that can be extracted from deep neural network models, as applying multi 1D convolution layers to input pattern. To evaluate performance of the proposed neural network model, as a result of comparing the previous recurrent neural network, our method showed a remarkable 98.4% accuracy.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.11
• /
• pp.3493-3506
• /
• 2022

In the field of scene segmentation, the precise segmentation of object boundaries in sports movement scene images is a great challenge. The geometric information and spatial information of the image are very important, but in many models, they are usually easy to be lost, which has a big influence on the performance of the model. To alleviate this problem, a parallel dense dilated convolution merging Network (termed PDDCM-Net) was proposed. The proposed PDDCMNet consists of a feature extractor, parallel dilated convolutions, and dense dilated convolutions merged with different dilation rates. We utilize different combinations of dilated convolutions that expand the receptive field of the model with fewer parameters than other advanced methods. Importantly, PDDCM-Net fuses both low-level and high-level information, in effect alleviating the problem of accurately segmenting the edge of the object and positioning the object position accurately. Experimental results validate that the proposed PDDCM-Net achieves a great improvement compared to several representative models on the COCO-Stuff data set.

• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.

• Journal of IKEEE
• /
• v.21 no.4
• /
• pp.358-367
• /
• 2017

Advanced Driver Assistance Systems(ADAS), such as Front Collision Warning System (FCWS) are currently being developed. FCWS require high processing speed because it must operate in real time while driving. In addition, a low-power system is required to operate in an automobile embedded system. In this paper, FCWS is implemented in CPU-FPGA architecture in embedded system to enable real-time processing. The lane detection enabled the use of the Inverse Transform Perspective (IPM) and sliding window methods to operate at fast speed. To detect the vehicle, a Convolutional Neural Network (CNN) with high recognition rate and accelerated by parallel processing in FPGA is used. The proposed architecture was verified using Intel FPGA Cyclone V SoC(System on Chip) with ARM-Core A9 which operates in low power and on-board FPGA. The performance of FCWS in HD resolution is 44FPS, which is real time, and energy efficiency is about 3.33 times higher than that of high performance PC enviroment.

• International Journal of Internet, Broadcasting and Communication
• /
• v.9 no.4
• /
• pp.1-7
• /
• 2017

Clipart is artificial visual contents that are created using various tools such as Illustrator to highlight some information. Here, the style of the clipart plays a critical role in determining how it looks. However, previous studies on clipart are focused only on the object recognition [16], segmentation, and retrieval of clipart images using hand-craft image features. Recently, some clipart classification researches based on the style similarity using CNN have been proposed, however, they have used different CNN-models and experimented with different benchmark dataset so that it is very hard to compare their performances. This paper presents an experimental analysis of the clipart classification based on the style similarity with two well-known CNN-models (Inception Resnet V2 [13] and VGG-16 [14] and transfers learning with the same benchmark dataset (Microsoft Style Dataset 3.6K). From this experiment, we find out that the accuracy of Inception Resnet V2 is better than VGG for clipart style classification because of its deep nature and convolution map with various sizes in parallel. We also find out that the end-to-end training can improve the accuracy more than 20% in both CNN models.

• Journal of Platform Technology
• /
• v.12 no.2
• /
• pp.58-63
• /
• 2024

CNNs tend to take a long time to learn and consume a lot of power due to lack of system resources with many data processing units when there are repetitive handles that do not have high performance in the image field. In this paper, we propose a handling method based on a low-power bus that can increase the exchange rate of multipliers and multiplicands within the convolution mixer, which is a tendency activity that occurs when a convolution mixer has multiplication, which is the core element of combination. Convolutional neural networks have proprietary low-power shared processor support and the design was implemented on an Intel DE1-SoC FPGA board using Verilog-HDL. The experiments validated the performance by comparing it with the exchange rate of the multiplier originally proposed by Shen on MNIST's numeric image database.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.3
• /
• pp.401-406
• /
• 2021

Recently, mean-teacher models based on convolutional recurrent neural networks are popularly used in audio event detection. The mean-teacher model is an architecture that consists of two parallel CRNNs and it is possible to train them effectively on the weakly-labelled and unlabeled audio data by using the consistency learning metric at the output of the two neural networks. In this study, we tried to improve the performance of the mean-teacher model by using additional derivative features of the log-mel spectrum. In the audio event detection experiments using the training and test data from the Task 4 of the DCASE 2018/2019 Challenges, we could obtain maximally a 8.1% relative decrease in the ER(Error Rate) in the mean-teacher model using proposed derivative features.