• Journal of information and communication convergence engineering
• /
• v.21 no.4
• /
• pp.346-350
• /
• 2023

Trajectory prediction is an essential element for driving autonomous vehicles, and various trajectory prediction models have emerged with the development of deep learning technology. Convolutional neural network (CNN) is the most commonly used neural network architecture for extracting the features of visual images, and the latest models exhibit high performances. This study was conducted to identify an efficient CNN backbone model among the components of deep learning models for trajectory prediction. We changed the existing CNN backbone network of multiple-trajectory prediction models used as feature extractors to various state-of-the-art CNN models. The experiment was conducted using nuScenes, which is a dataset used for the development of autonomous vehicles. The results of each model were compared using frequently used evaluation metrics for trajectory prediction. Analyzing the impact of the backbone can improve the performance of the trajectory prediction task. Investigating the influence of the backbone on multiple deep learning models can be a future challenge.

• Journal of Engineering Education Research
• /
• v.17 no.6
• /
• pp.77-83
• /
• 2014

This paper deals with the practical experience and lessons to develop the criteria for evaluating projects in Computer Science & Information Engineering. In engineering education, project-based learning is different from problem-based learning. Project tasks are closer to professional reality. And self-direction is stronger in project work. In this paper, the development of project evaluation criteria is introduced with consideration of these differences. It is explained through running examples of basic project courses in Computer Science & Information Engineering.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06a
• /
• pp.299-300
• /
• 2010

• Annual Conference of KIPS
• /
• 2017.04a
• /
• pp.928-931
• /
• 2017

네트워크 트래픽 데이터를 정제하여, Convolutional Neural Network Model 훈련에 적합한 데이터 세트로 변환하는데, 그 방법은 패킷 단위의 트래픽 데이터를 이미지 형태로 만드는 것이다. 완성된 데이터 세트를 훈련데이터로 하여 Convolutional Neural Network Model에 훈련하고, 훈련데이터의 이미지 크기를 변환해가며 훈련시킨 결과에 대해 비교 분석 및 평가를 진행한다.

• Annual Conference of KIPS
• /
• 2016.10a
• /
• pp.628-631
• /
• 2016

'야구경기 결과에 대해 자동으로 기사를 작성할 수 있는가'에서 본 논문에서는 야구 경기 데이터들을 기반으로 의사결정 트리기법을 사용하여 경기결과의 문맥과 기사작성에 필요한 요소들을 자동으로 추출해보았다. 그 결과 해당경기의 데이터를 가지고 객관적인 야구기사를 생산해 낼 수 있음을 도출해냈다.

• Annual Conference of KIPS
• /
• 2011.11a
• /
• pp.277-278
• /
• 2011

본 논문에서는 공압출 다층 필름 제조공정에서 수지의 품질에 영향을 주는 외관상의 결함을 검사하는 비젼 시스템을 제안한다. 필름 생산 과정에서는 흑점이나 주름 등을 포함한 다양한 결함이 발생할 수 있는데, 명암이 명확히 구별되는 결함도 있지만 그렇지 않은 결함들은 필름의 특성에 의해 검출 및 분류가 어려운 경우가 많다. 제안된 논문에서는 전체 검사시스템의 소개와 함께 결함의 종류 분류와 검출 및 분류 방법을 제안하는데, 특히 애매한 결함의 구분을 위해 지역적 이진패턴(LBP)에 기반한 텍스쳐 분석 방법을 이용한다. 실험을 통해 제안된 시스템 및 방법이 필름 생산과정의 다양한 결함들을 잘 검출하고 분류하는 것을 알 수 있었다.

• Journal of Engineering Education Research
• /
• v.17 no.2
• /
• pp.68-74
• /
• 2014

Game industry demands for talented workers from game scenario writing, graphic design and computer game programming areas, which are the three heterogeneous parts of computer game making. This paper deals with the analysis on the college convergence education, surveying the curriculums of game departments in Korean universities and with how to proceed for game programmers to the direction of convergence education for computer game making. Most of the game departments are focusing on game programming area, whereas only 2 departments among 14 surveyed are teaching the game graphic design topics. Observing the curriculum changes for sampled game departments reveals that around 30% of the majoring classes are directly game related. Game programmers are to be advised to fulfill their programming skill first and later expand their ability to game graphic design skill.

• Journal of Practical Engineering Education
• /
• v.14 no.2
• /
• pp.439-445
• /
• 2022

Due to the development of online media platforms and the COVID-19 incident, the mass production and consumption of digital video content are rapidly increasing. In order to select digital video content, users grasp it in a short time through thumbnails and teaser videos, and select and watch digital video content that suits them. It is very inconvenient to check all digital video contents produced around the world one by one and manually edit teaser videos for users to choose from. In this paper, keyframes are extracted based on YCbCr color models to automatically generate teaser videos, and keyframes extracted through clustering are optimized. Finally, we present a method of producing a teaser video to help users check digital video content by connecting the finally extracted keyframes.

• Journal of information and communication convergence engineering
• /
• v.8 no.2
• /
• pp.191-196
• /
• 2010

This paper presents automatic generation of robot motion using posture template in physical training system. In this system, we define a human-like generic template with two angles, then, we generate robot motion that has used angle data of the created template. The angle data is transformed automatically to the most similar motion. This is achieved by suitable expressions for robot specification. The expressions intend to differences of angle measurement on X,Y,Z planes and DOF constraint. We choose several postures and create generic templates for chosen postures. Robot motions are also generated from generic templates by expressions.

• Journal of information and communication convergence engineering
• /
• v.8 no.2
• /
• pp.180-184
• /
• 2010

This paper proposes a robot motion authoring using finger-robot interaction. The proposed method is a user friendly method that easily authors (creates and controls) robot motion according to the number of fingers. The effectiveness of the proposed motion authoring method was verified based on motion authoring simulation of an industrial robot.