• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.927-929
• /
• 2021

최근 펫팸족(Pet-Family)과 같이 반려동물을 가족처럼 생각하는 가구가 증가하면서 반려동물 시장이 크게 성장하고 있다. 이러한 이유로 본 논문에서는 반려동물의 객체 식별을 통한 객체 분할과 신체 좌표추정에 기반을 둔 반려동물의 행동 분류 방법을 제안한다. 이 방법은 CCTV를 통해 반려동물 영상 데이터를 수집한다. 수집된 영상 데이터는 반려동물의 인스턴스 분할을 위해 Mask R-CNN(Region Convolutional Neural Networks) 모델을 적용하고, DeepLabCut 모델을 통해 추정된 신체 좌푯값을 도출한다. 이 결과로 도출된 영상 데이터와 추정된 신체 좌표 값은 CNN(Convolutional Neural Networks)-LSTM(Long Short-Term Memory) 모델을 적용하여 행동을 분류한다. 본 모델을 바탕으로 행동을 분석 및 분류하여, 반려동물의 위험 상황과 돌발 행동에 대한 올바른 대처를 제공할 수 있는 기반을 제공할 것이라 기대한다.

• Korean Journal of Remote Sensing
• /
• v.37 no.5_1
• /
• pp.1083-1094
• /
• 2021

In order to monitor nuclear activity in inaccessible areas, it is necessary to establish a methodology to analyze changesin nuclear activity-related objects using high-resolution satellite images. However, traditional object detection and change detection techniques using satellite images have difficulties in applying detection results to various fields because effects of seasons and weather at the time of image acquisition. Therefore, in this paper, an object of interest was detected in a satellite image using a deep learning model, and object changes in the satellite image were analyzed based on object detection results. An initial training of the deep learning model was performed using an open dataset for object detection, and additional training dataset for the region of interest were generated and applied to transfer learning. After detecting objects by multitemporal and multisensory satellite images, we tried to detect changes in objects in the images by using them. In the experiments, it was confirmed that the object detection results of various satellite images can be directly used for change detection for nuclear activity-related monitoring in inaccessible areas.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.2
• /
• pp.231-237
• /
• 2020

In this paper, we provide an efficient hybrid video blending scheme to improve the naturalness of composite video in Poisson equation-based composite methods. In image blending process, various blending methods are used depending on the purpose of image composition. The hybrid blending method proposed in this paper has the characteristics that there is no seam in the composite video and the color distortion of the object is reduced by properly utilizing the advantages of Poisson blending and alpha blending. First, after blending the source object by the Poisson blending method, the color difference between the blended object and the original object is compared. If the color difference is equal to or greater than the threshold value, the object of source video is alpha blended and is added together with the Poisson blended object. Simulation results show that the proposed method has not only better naturalness than Poisson blending and alpha blending, but also requires a relatively small amount of computation.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.8
• /
• pp.1919-1925
• /
• 2013

This paper proposes a background image extraction algorithm for traffic environment analysis in a school zone. The proposed algorithm solves the problems by level changes and stationary objects to be occurred frequently in traffic environment. For the former, it renews rapidly the background image toward the current frame using a fast Sima-Delta algorithm and for the latter, it excludes the stationary objects from the background image by detecting dynamic regions using a just previous frame and a background image averaged for a long time. The results of experiments show that the proposed algorithm adapts quickly itself to level change well, and reduces about 40~80% of SAD in background region in comparison with the conventional algorithms.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.5
• /
• pp.13-20
• /
• 2019

Recently, object detection technology based on CNN has been actively studied. Object detection technology is used as an important technology in autonomous vehicles, intelligent image analysis, and so on. In this paper, we propose a rotation change robust object detection system by applying MI-FL (Moment Invariant-Feature Layer) to SSD (Single Shot Multibox Detector) which is one of CNN-based object detectors. First, the features of the input image are extracted based on the VGG network. Then, a total of six feature layers are applied to generate bounding boxes by predicting the location and type of object. We then use the NMS algorithm to get the bounding box that is the most likely object. Once an object bounding box has been determined, the invariant moment feature of the corresponding region is extracted using MI-FL, and stored and learned in advance. In the detection process, it is possible to detect the rotated image more robust than the conventional method by using the previously stored moment invariant feature information. The performance improvement of about 4 ~ 5% was confirmed by comparing SSD with existing SSD and MI-FL.

• Journal of the Korea Computer Graphics Society
• /
• v.2 no.1
• /
• pp.77-85
• /
• 1996

Using a series of medical tomograms, we can reconstruct internal organs or other objects of interest and generate 3-D images. It is generally accepted that the axial resolution determined by two sequential image slices is lower than the planar resolution in one image slices. Therefore, various methods of interpolation were developed for an accurate display of reconstructed images. In this paper, a new algorithm for 3-D reconstruction of the medical images such as MRI and X-ray CT is suggested. The algorithm is shape-based and utilizes parts of the gray-level information. We extend the conventional shape-based interpolation of the binary images to the gray-scale images using the shortest distance map. Using this new algorithm, We could reduce the execution time for interpolation while keeping similar high quality of the reconstructed images with reduced execution time and is applicable to the various medical tomograms.

• The KIPS Transactions:PartB
• /
• v.8B no.5
• /
• pp.493-500
• /
• 2001

Video scene segmentation is fundamental role for content based video analysis. In this paper, we propose a new region based video scene segmentation method using continuity test for each object region which is segmented by the watershed algorithm for all frames in video data. For this purpose, we first classify video data segments into classes that are the dynamic and static sections according to the object movement rate by comparing the spatial and shape similarity of each region. And then, try to segment each sections by grouping each sections by comparing the neighbor section sections by comparing the neighbor section similarity. Because, this method uses the region which represented on object as a similarity measure, it can segment video scenes efficiently without undesirable fault alarms by illumination and partial changes.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.3
• /
• pp.101-108
• /
• 2019

Due to recent improvements in computer processing speed and image processing technology, researches are being actively carried out to combine information from camera with existing GNSS (Global Navigation Satellite System) and dead reckoning. In this study, developed a vision-based positioning assistant algorithm to estimate the distance to the object from stereo images. In addition, GNSS/on-board vehicle sensor/vision based positioning algorithm is developed by combining vision based positioning algorithm with existing positioning algorithm. For the performance analysis, the velocity calculated from the actual driving test was used for the navigation solution correction, simulation tests were performed to analyse the effects of velocity precision. As a result of analysis, it is confirmed that about 4% of position accuracy is improved when vision information is added compared to existing GNSS/on-board based positioning algorithm.

• Journal of Korea Multimedia Society
• /
• v.5 no.2
• /
• pp.167-175
• /
• 2002

Image Mosaicing creates a new image by composing video frames or still images that are related, and performed by arrangement, composition and redundancy analysis of images. This paper proposes a hierarchical image mosaicing system using camera and object parameters far efficient video database construction. A tree-based image mosiacing has implemented for high-speed computation time and for construction of static and dynamic image mosaic. Camera parameters are measured by using least sum of squared difference and affine model. Dynamic object detection algorithm has proposed for extracting dynamic objects. For object extraction, difference image, macro block, region splitting and 4-split detection methods are proposed and used. Also, a dynamic positioning method is used for presenting dynamic objects and a blurring method is used for creating flexible mosaic image.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.9C
• /
• pp.555-563
• /
• 2011

In this paper, we propose an object based 2D image to 3D conversion algorithm by using motion estimation, color labeling and non-local mean filtering methods. In the proposed algorithm, we first extract the motion vector of each object by estimating the motion between frames and then segment a given image frame with color labeling method. Then, combining the results of motion estimation and color labeling, we extract object regions and assign an exact depth value to each object to generate the right image. While generating the right image, occlusion regions occur but they are effectively recovered by using non-local mean filter. Through the experimental results, it is shown that the proposed algorithm performs much better than conventional conversion scheme by removing the eye fatigue effectively.