• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.1
• /
• pp.13-20
• /
• 2016

In this paper, we propose an object recognition system that can effectively find out its category, its instance name, and several attributes from the color and depth images of an object with hierarchical feature learning. In the preprocessing stage, our system transforms the depth images of the object into the surface normal vectors, which can represent the shape information of the object more precisely. In the feature learning stage, it extracts a set of patch features and image features from a pair of the color image and the surface normal vector through two-layered learning. And then the system trains a set of independent classification models with a set of labeled feature vectors and the SVM learning algorithm. Through experiments with UW RGB-D Object Dataset, we verify the performance of the proposed object recognition system.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.11 no.3
• /
• pp.129-135
• /
• 2011

In this paper, we propose the algorithm using optical flow and machine learning-based segmentation for the 3D conversion of 2D video. For the segmentation allowing the successful 3D conversion, we design a new energy function, where color/texture features are included through machine learning method and the optical flow is also introduced in order to focus on the regions with the motion. The depth map are then calculated according to the optical flow of segmented regions, and left/right images for the 3D conversion are produced. Experiment on various video shows that the proposed method yields the reliable segmentation result and depth map for the 3D conversion of 2D video.

• Journal of Korea Multimedia Society
• /
• v.5 no.5
• /
• pp.505-514
• /
• 2002

In this Paper, we propose a method of content-based image classification using neural network. The images for classification ate object images that can be divided into foreground and background. To deal with the object images efficiently, object region is extracted with a region segmentation technique in the preprocessing step. Features for the classification are texture and shape features extracted from wavelet transformed image. The neural network classifier is constructed with the extracted features and the back-propagation learning algorithm. Among the various texture features, the diagonal moment was more effective. A test with 300 training data and 300 test data composed of 10 images from each of 30 classes shows correct classification rates of 72.3% and 67%, respectively.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.339-342
• /
• 2021

In this paper, we propose a new method for inactive region padding using reinforcement learning. Inactive region is an area that has no information, such as 360 or 3DOF+ vidoes. However, these inactive regions degrade the compression performance in general. To improve the compression performance, simple filtering is applied between active and inactive regions. But it does not fully consider the characteristics of the images. In the proposed method, inactive regions are padded through reinforcement learning that can consider the characteristics of images and the compression process. Experimental results show that the performance is better than the conventional padding method.

• The Journal of the Convergence on Culture Technology
• /
• v.10 no.3
• /
• pp.909-914
• /
• 2024

This study suggests a quantitative evaluation of transfer learning, which is widely used in various AI fields, including image recognition for robot vision. Quantitative and qualitative analyses of results applying transfer learning are presented, but transfer learning itself is not discussed. Therefore, this study proposes a quantitative evaluation of transfer learning itself based on MNIST, a handwritten digit database. For the reference network, the change in recognition accuracy according to the depth of the transfer learning frozen layer and the ratio of transfer learning data and pre-training data is tracked. It is observed that when freezing up to the first layer and the ratio of transfer learning data is more than 3%, the recognition accuracy of more than 90% can be stably maintained. The transfer learning quantitative evaluation method of this study can be used to implement transfer learning optimized according to the network structure and type of data in the future, and will expand the scope of the use of robot vision and image analysis AI in various environments.

• Journal of Internet Computing and Services
• /
• v.8 no.3
• /
• pp.19-32
• /
• 2007

This paper present a new face recognition method based on LVQ neural net to construct a real time face recognition system. The previous researches which used PCA, LDA combined neural net usually need much time in training neural net. The supervised LVQ neural net needs much less time in training and can maximize the separability between the classes. In this paper, the proposed method transforms the input face image by PCA and LDA sequentially into low-dimension feature vectors and recognizes the face through LVQ neural net. In order to make the system robust to external light variation, light compensation is performed on the detected face by max-min normalization method as preprocessing. PCA and LDA transformations are applied to the normalized face image to produce low-level feature vectors of the image. In order to determine the initial centers of LVQ and speed up the convergency of the LVQ neural net, the K-Means clustering algorithm is adopted. Subsequently, the class representative vectors can be produced by LVQ2 training using initial center vectors. The face recognition is achieved by using the euclidean distance measure between the center vector of classes and the feature vector of input image. From the experiments, we can prove that the proposed method is more effective in the recognition ratio for the cases of still images from ORL database and sequential images rather than using conventional PCA of a hybrid method with PCA and LDA.

• The Journal of the Korea Contents Association
• /
• v.9 no.2
• /
• pp.453-461
• /
• 2009

In this paper, we design and implement a web-based courseware where learners can do self-directed learning to study Web programming languages such as Html, CSS, JavaScript and Dhtml. Each section consists of text class, movie class, practice class, formative assessment, laboratory and bulletin board. And our courseware makes teachers to teach, assess and give scores to learners on Web. In our Web courseware, learners can play a movie class and practice Web programming codes in one screen simultaneously, and execute codes and confirm a results in the same screen also. Therefore, learners can understand Web programming languages efficiently, which are hard to understand immediately by text.

• Proceedings of the Korea Contents Association Conference
• /
• 2012.05a
• /
• pp.383-384
• /
• 2012

본 연구는 온라인 구성주의 학습 등 다양한 체험학습이 가능하게 하기 위하여 웹캠으로 촬영된 영상 중 사용자의 모습만을 추출, 이를 전송하여 여러 사용자의 영상과 교육과정에 맞는 상황이나 현장을 3D 환경에 혼합하여 제공하는 시스템을 개발하는데 있어 영상, 음성 3D 정보, 학습개체 등을 원활하게 통신할 수 있는 통신 매커니즘을 설계하고 구현한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.04b
• /
• pp.691-693
• /
• 2002

현대 사회에서 수요자를 고려하지 않은 산업 형태는 거의 찾아볼 수 없다. 우리 교육 또한 학생 개개인의 능력을 출발점으로 하여 학생의 눈높이에 맞춘 교육을 바람직하게 여기는 추세로 변화하고 있다. 이렇게 볼 때 가장 이상적인 학습형태는 개별화 수업이다. 최근 데이터 압축 기술과 통신 기술의 발달로 동영상 데이터(video data)를 이용한 다양한 서비스가 가능하게 되면서 교육에서도 그 사용의 폭이 넓어지고 있다. 이에 따라 동영상 데이터를 효율적으로 저장, 관리할 수 있는 검색에 대한 연구가 활발하게 이루어져오고 있다. 본 논문에서는 동영상 데이터 검색방법의 하나인 주석기반 방식(text-based retrieval)을 이용하여 하나의 교육용 비디오를 장면분할(scene segmentation)하여 학습내용의 수준에 따라 상.중.하 색인을 한다. 이 색인된 비디오에서 수준별 개별학습이 가능한 가상의 비디오 시퀸스(video sequence)를 만들어낸다.

• Proceedings of the Korea Inteligent Information System Society Conference
• /
• 2004.11a
• /
• pp.195-202
• /
• 2004

일반적으로 운송 컨테이너의 식별자들은 크기나 위치가 정형화되어 있지 않고 외부 잡음으로 인하여 식별자의 형태가 변형될 수 있기 때문에 일정한 규칙으로 찾기는 힘들다. 본 논문에서는 이러한 특성을 고려하여 컨테이너 영상에 대해 Canny 마스크를 이용하여 에지를 검출하고, 검출된 에지 정보에서 영상획득 시 외부 광원에 의해 수직으로 길게 발생하는 잡음들을 퍼지추론 방법을 적용하여 제거한 후에 수직 블록과 수평 블록을 검출하여 컨테이너의 식별자 영역을 추출하고 이진화 한다. 이진화된 식별자 영역에 대해 검정색의 빈도수를 이용하여 흰바탕과 민바탕을 구분하고 윤곽선 추적 알고리즘을 적용하여 개별 식별자를 추출한다. 추출된 개별 식별자의 인식은 개선된 퍼지 RBF 네트워크를 제안하여 적용한다. 제안된 퍼지 RBF 네트워크는 퍼지 C-Means 알고리즘을 중간층으로 적용하고 중간층과 출력층 간의 학습에는 일반화된 델타 학습 방법과 Delta-bar-Delta 알고리즘을 적용하여 학습 성능을 개선한다. 실제 컨테이너 영상을 대상으로 실험한 결과, 기존의 식별자 추출 방법보다 제안된 식별자 추출 방법이 개선되었고 기존의 퍼지 RBF 네트워크 보다 제안된 퍼지 RBF 네트워크가 컨테이너 식별자의 학습 및 인식에 우수함을 확인하였다.