• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.1
• /
• pp.149-156
• /
• 2024

The increased intelligence and speed of vehicle infotainment, whose main purpose was emergency and external communication, is showing the potential for application to various services such as navigation and autonomous driving. In particular, functionality for linking external devices and infrastructure is being strengthened due to advances in communication and networks. Under this trend, it is necessary to consider the direction of linkage with the cooperative intelligent transportation system (C-ITS) for advanced vehicle services and driving. In addition, in the case of automobiles, future vehicle development concepts are being established based on the concept of software-defined vehicles (SDVs) in line with the trend of electrification beyond telematics and infotainment advancements, and such SDV linkage must be considered at the same time. In this paper, we consider the future direction of ITS and SDV linkage based on analysis of vehicle infotainment structure. First, for this purpose, we analyze the existing vehicle infotainment structure and architecture, and also present the structure of the SDV linked to it. Based on this, analysis and implications are drawn on the possibility of applying and linking standard-based C-ITS services with SDV devices.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.9
• /
• pp.375-384
• /
• 2021

Deep learning techniques have been proven to have high performance in image processing and are applied in various fields. The most widely used methods for validating a deep learning model include a holdout verification method, a k-fold cross verification method, and a bootstrap method. These legacy methods consider the balance of the ratio between classes in the process of dividing the data set, but do not consider the ratio of various features that exist within the same class. If these features are not considered, verification results may be biased toward some features. Therefore, we propose a deep learning model validation method based on data feature coverage for image classification by improving the legacy methods. The proposed technique proposes a data feature coverage that can be measured numerically how much the training data set for training and validation of the deep learning model and the evaluation data set reflects the features of the entire data set. In this method, the data set can be divided by ensuring coverage to include all features of the entire data set, and the evaluation result of the model can be analyzed in units of feature clusters. As a result, by providing feature cluster information for the evaluation result of the trained model, feature information of data that affects the trained model can be provided.

• Journal of Practical Engineering Education
• /
• v.15 no.2
• /
• pp.529-538
• /
• 2023

At the present time when the need for universal artificial intelligence education is expanding and job changes are being made, research and discussion on artificial intelligence liberal arts education for non-majors in universities who experience artificial intelligence as part of their job is insufficient. Although artificial intelligence education courses for non-majors are being operated, they are mainly operated as theory-oriented education on the concepts and principles of artificial intelligence. In order to understand the general concept of artificial intelligence for non-majors, it is necessary to proceed with experiential learning in parallel. Therefore, this study designs artificial intelligence experiential education learning contents of difficulty that can reduce the burden of artificial intelligence classes with interest in learning by considering the characteristics of non-majors. After, we will examine the learning effect of experiential education using App Inventor and the Orange artificial intelligence platform. As a result of analysis based on the learning-related data and survey data collected through the creation of AI-related projects by teams, positive changes in the perception of the need for AI education were found, and AI literacy skills improved. It is expected that it will serve as an opportunity for instructors to lay the groundwork for designing a learning model for artificial intelligence experiential education learning.

• Korean Journal of Radiology
• /
• v.24 no.4
• /
• pp.294-304
• /
• 2023

Objective: We aimed to investigate whether image standardization using deep learning-based computed tomography (CT) image conversion would improve the performance of deep learning-based automated hepatic segmentation across various reconstruction methods. Materials and Methods: We collected contrast-enhanced dual-energy CT of the abdomen that was obtained using various reconstruction methods, including filtered back projection, iterative reconstruction, optimum contrast, and monoenergetic images with 40, 60, and 80 keV. A deep learning based image conversion algorithm was developed to standardize the CT images using 142 CT examinations (128 for training and 14 for tuning). A separate set of 43 CT examinations from 42 patients (mean age, 10.1 years) was used as the test data. A commercial software program (MEDIP PRO v2.0.0.0, MEDICALIP Co. Ltd.) based on 2D U-NET was used to create liver segmentation masks with liver volume. The original 80 keV images were used as the ground truth. We used the paired t-test to compare the segmentation performance in the Dice similarity coefficient (DSC) and difference ratio of the liver volume relative to the ground truth volume before and after image standardization. The concordance correlation coefficient (CCC) was used to assess the agreement between the segmented liver volume and ground-truth volume. Results: The original CT images showed variable and poor segmentation performances. The standardized images achieved significantly higher DSCs for liver segmentation than the original images (DSC [original, 5.40%-91.27%] vs. [standardized, 93.16%-96.74%], all P < 0.001). The difference ratio of liver volume also decreased significantly after image conversion (original, 9.84%-91.37% vs. standardized, 1.99%-4.41%). In all protocols, CCCs improved after image conversion (original, -0.006-0.964 vs. standardized, 0.990-0.998). Conclusion: Deep learning-based CT image standardization can improve the performance of automated hepatic segmentation using CT images reconstructed using various methods. Deep learning-based CT image conversion may have the potential to improve the generalizability of the segmentation network.

• Journal of the Korea Computer Graphics Society
• /
• v.25 no.3
• /
• pp.115-122
• /
• 2019

We In this study, We tested for artificial intelligence-based virtual space story books were planned. The proposed virtual space concept, a story book with the characteristics of Augmented Virtuality, was implemented Several steps are needed to proceed with the virtual space storybook's story. First, a user brings a real object in to virtual space and recognizes the real object with an artificial intelligence-based object-recognition software. Second, when object recognition progresses, the virtual 3D model is augmented in the virtual space, which is then inserted into the virtual space and rendered. Finally, software projected a virtual space storybook on the desk in which users can touch and select real-objects. This virtual space storybook was implemented using the new story-making technology by applying the virtual space concept. the Augmented Virtuality concept is to augment real objects based on virtual space. To confirm this we tested a user test using the virtual space storybook. the user seemed as if can the distinction between real objects and virtual images. Also very well and that understood the process of putting the real objects in virtual space.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.2
• /
• pp.173-178
• /
• 2019

Recently, the types and amount of data generated, collected, and measured in IoT such as smart home, security, and factory are increasing. The technologies for IoT service include sensor devices to measure desired data, embedded software to control the devices such as signal processing, wireless network protocol to transmit and receive the measured data, and big data and AI-based analysis. In this paper, we focused on developing a gateway for interfacing heterogeneous sensor network protocols that are used in various IoT devices and propose a heterogeneous network interface IoT gateway. We utilized a OpenWrt-based wireless routers and used 6LoWAN stack for IP-based communication via BLE and IEEE 802.15.4 adapters. We developed a software to convert Z-Wave and LoRa packets into IP packet using our Python-based middleware. We expect the IoT gateway to be used as an effective device for collecting IoT big data.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.11
• /
• pp.1642-1648
• /
• 2021

We analyzes representative technologies of IoT patents and reflects these results in the curriculum of IoT. In order to identify the representative technologies, the IPC codes of the patents were analyzed. Among the main category IPC codes, the most used IPC codes were H04L in Single IPC Patent with 974 cases(32.0%) and G06Q in Multiple IPC Patent with 710 cases(29.2%). As a result of classifying the IPC code into the WIPO technology classification system, the most emphasized technologies are Digital Communication, accounting for about 60.5% in the Single IPC Patent and IT Methods for Management(710 cases, 29.2%) in Multiple IPC Patent. The main points to be considered when organizing the curriculum of IoT are: ∇Emphasis on Digital Communication, ∇Expansion of Education related to IT Methods for Management(Including entrepreneurship and patent application), and ∇Consideration of subjects related to the Convergence of IoT. This research can contribute to the curriculum design of new industrial technologies such as AI and Fintech.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.4
• /
• pp.256-263
• /
• 2022

Skin cancer is one of the most common diseases in the world, and the incidence rate in Korea has increased by about 100% over the past five years. In the United States, more than 5 million people are diagnosed with skin cancer every year. Skin cancer mainly occurs when skin tissue is damaged for a long time due to exposure to ultraviolet rays. Melanoma, a malignant tumor of skin cancer, is similar in appearance to Atypical melanocytic nevus occurring on the skin, making it difficult for the general public to be aware of it unless secondary signs occur. In this paper, we propose a skin cancer lesion edge detection algorithm and a deep learning model, CRNN, which performs skin cancer lesion classification for early detection and classification of these skin cancers. As a result of the experiment, when using the contour detection algorithm proposed in this paper, the classification accuracy was the highest at 97%. For the Canny algorithm, 78% was shown, 55% for Sobel, and 46% for Laplacian.

• 한국정보교육학회:학술대회논문집
• /
• 2021.08a
• /
• pp.261-267
• /
• 2021

This study examines the purpose and goals of ICT education, compares them with current software and artificial intelligence-oriented information curriculum, analyzes foreign SW curriculum, extracts learning topics and elements, and analyzes whether the current information curriculum is presented effectively. As a result of the analysis, the number of information-related courses in Korea is currently lower than in other countries, which has reduced the number of basic computer applications and underlying software applications such as presentations and spreadsheets covered in ICT training in the past. In addition, compared to Korea's curriculum where information education begins in the fifth grade of elementary school, other countries are introducing information education from the first grade to the third grade of elementary school. Therefore, active discussions on the expansion of the number of information education, the timing of introduction of information education, and the utilization of basic computers are needed.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.879-884
• /
• 2004

This paper describes the development of a synthetic multi-agent called KMITL Cadence 2003. KMITL Cadence 2003 is a robotic soccer simulation team consisting of eleven autonomous software agents. Each agent operates in a physical soccer simulation model called Robocup Soccer Server which provides fully distributed and real-time multi-agent system environment. All teammates have to cooperate to achieve the common goal of winning the game. The simulation models many aspects of the football field such as noise in ball movements, noisy sensors, unreliable communication channel between teammates and actuators, limited physical abilities and restricted communication. This paper addresses the algorithm to develop the soccer agents to perform basic actions which are scoring, passing ball and blocking the opponents effectively. The result of this development is satisfactory because the successful scoring attempts is increased from 11.1% to 33.3%, successful passing ball attempts is increased from 22.08% to 63.64%, and also, successful intercepting attempts is increased from 88% to 97.73%.