• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.4 s.310
• /
• pp.74-86
• /
• 2006

This paper proposes a robust and computationally efficient algorithm for automatic video object segmentation. For implementing the spatio-temporal segmentation, which aims for efficient combination of the motion segmentation and the color segmentation, an SOM-based hierarchical clustering method in which the segmentation process is regarded as clustering of feature vectors is employed. As results, problems of high computational complexity which required for obtaining exact segmentation results in conventional video object segmentation methods, and the performance degradation due to noise are significantly reduced. A measure of motion vector reliability which employs MRF-based MAP estimation scheme has been introduced to minimize the influence from the motion estimation error. In addition, a noise elimination scheme based on the motion reliability histogram and a clustering validity index for automatically identifying the number of objects in the scene have been applied. A cross projection method for effective object tracking and a dynamic memory to maintain temporal coherency have been introduced as well. A set of experiments has been conducted over several video sequences to evaluate the proposed algorithm, and the efficiency in terms of computational complexity, robustness from noise, and higher segmentation accuracy of the proposed algorithm have been proved.

• Journal of Intelligence and Information Systems
• /
• v.28 no.1
• /
• pp.89-106
• /
• 2022

Recently, the amount of video data collected from smartphones, CCTVs, black boxes, and high-definition cameras has increased rapidly. According to the increasing video data, the requirements for analysis and utilization are increasing. Due to the lack of skilled manpower to analyze videos in many industries, machine learning and artificial intelligence are actively used to assist manpower. In this situation, the demand for various computer vision technologies such as object detection and tracking, action detection, emotion detection, and Re-ID also increased rapidly. However, the object detection and tracking technology has many difficulties that degrade performance, such as re-appearance after the object's departure from the video recording location, and occlusion. Accordingly, action and emotion detection models based on object detection and tracking models also have difficulties in extracting data for each object. In addition, deep learning architectures consist of various models suffer from performance degradation due to bottlenects and lack of optimization. In this study, we propose an video analysis system consists of YOLOv5 based DeepSORT object tracking model, SlowFast based action recognition model, Torchreid based Re-ID model, and AWS Rekognition which is emotion recognition service. Proposed model uses single-linkage hierarchical clustering based Re-ID and some processing method which maximize hardware throughput. It has higher accuracy than the performance of the re-identification model using simple metrics, near real-time processing performance, and prevents tracking failure due to object departure and re-emergence, occlusion, etc. By continuously linking the action and facial emotion detection results of each object to the same object, it is possible to efficiently analyze videos. The re-identification model extracts a feature vector from the bounding box of object image detected by the object tracking model for each frame, and applies the single-linkage hierarchical clustering from the past frame using the extracted feature vectors to identify the same object that failed to track. Through the above process, it is possible to re-track the same object that has failed to tracking in the case of re-appearance or occlusion after leaving the video location. As a result, action and facial emotion detection results of the newly recognized object due to the tracking fails can be linked to those of the object that appeared in the past. On the other hand, as a way to improve processing performance, we introduce Bounding Box Queue by Object and Feature Queue method that can reduce RAM memory requirements while maximizing GPU memory throughput. Also we introduce the IoF(Intersection over Face) algorithm that allows facial emotion recognized through AWS Rekognition to be linked with object tracking information. The academic significance of this study is that the two-stage re-identification model can have real-time performance even in a high-cost environment that performs action and facial emotion detection according to processing techniques without reducing the accuracy by using simple metrics to achieve real-time performance. The practical implication of this study is that in various industrial fields that require action and facial emotion detection but have many difficulties due to the fails in object tracking can analyze videos effectively through proposed model. Proposed model which has high accuracy of retrace and processing performance can be used in various fields such as intelligent monitoring, observation services and behavioral or psychological analysis services where the integration of tracking information and extracted metadata creates greate industrial and business value. In the future, in order to measure the object tracking performance more precisely, there is a need to conduct an experiment using the MOT Challenge dataset, which is data used by many international conferences. We will investigate the problem that the IoF algorithm cannot solve to develop an additional complementary algorithm. In addition, we plan to conduct additional research to apply this model to various fields' dataset related to intelligent video analysis.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.1
• /
• pp.120-129
• /
• 1990

In this paper, 3-dimensional object segmentation and classification are proposed. Planar object is segmented surface using jump boundary and internal boundary. Curved object is segmented surfaces by maximin clustering method. Segmented surfaces are classified by depth trends and angle measurement of normal vectors. Classified surfaces are merged according to adjacent surfaces and compared to Guassian curvature and mean curvature method. The proposed methods have been successfully applied to the synthetic range images and shows good classification.

• Proceedings of the IEEK Conference
• /
• 2002.06d
• /
• pp.141-144
• /
• 2002

This paper describes a new feature based tracking system that can track moving objects with a pan-tilt camera. We extract corner features of the scene and tracks the features using filtering, The global motion energy caused by camera movement is eliminated by finding the maximal matching position between consecutive frames using Pyramidal template matching. The region of moving object is segmented by clustering the motion trajectories and command the pan-tilt controller to follow the object such that the object will always lie at the center of the camera. The proposed system has demonstrated good performance for several video sequences.

• Journal of Intelligence and Information Systems
• /
• v.8 no.2
• /
• pp.177-187
• /
• 2002

Spatial data mining is a process to discover interesting relationships and characteristics those exist implicitly in a spatial database. Many spatial clustering algorithms have been developed. But, there are few approaches that focus simultaneously on clustering spatial data and assigning weight to non-spatial attributes of objects. In this paper, we propose a new spatial clustering algorithm, called DBSCAN-W, which is an extension of the existing density-based clustering algorithm DBSCAN. DBSCAN algorithm considers only the location of objects for clustering objects, whereas DBSCAN-W considers not only the location of each object but also its non-spatial attributes relevant to a given application. In DBSCAN-W, each datum has a region represented as a circle of various radius, where the radius means the degree of the importance of the object in the application. We showed that DBSCAN-W is effective in generating clusters reflecting the users requirements through experiments.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.4 no.2
• /
• pp.108-113
• /
• 2003

Research results have demonstrated that conventional client-server databases have scalability problem in the presence of many concurrent clients. The multi-tier architecture that exploits similarities in clients' object access behavior partitions clients into logical clusters according to their object request pattern. As a result, object requests that are served inside the clusters, server load and request response time can be optimized. Management of clustering by utilizing clients' access pattern-based is an important component for the system's goal. Off-line methods optimizes the quality of the global clustering, the necessary cost and clustering schedule needs to be considered and planned carefully in respect of stable system's performance. In this paper, we propose methods that detect changes in access behavior and optimize system configuration in real time. Finally this paper demonstrates the effectiveness of on-line change detection and results of experimental investigation concerning reconfiguration.

• 한국데이터정보과학회:학술대회논문집
• /
• 2003.10a
• /
• pp.249-258
• /
• 2003

K-means clustering has been widely used in many applications, such that pattern analysis or recognition, data analysis, image processing, market research and so on. It can identify dense and sparse regions among data attributes or object attributes. But k-means algorithm requires many hours to get k clusters that we want, because it is more primitive, explorative. In this paper we propose a new method of k-means clustering using the grid-based sample. It is more fast than any traditional clustering method and maintains its accuracy.

• 한국데이터정보과학회:학술대회논문집
• /
• 2003.10a
• /
• pp.229-238
• /
• 2003

K-means clustering has been widely used in many applications, such that pattern analysis, data analysis, market research and so on. It can identify dense and sparse regions among data attributes or object attributes. But k-means algorithm requires many hours to get k clusters, because it is more primitive and explorative. In this paper we propose a new method of k-means clustering using the grid-based representative value(arithmetic and trimmed mean) for sample. It is more fast than any traditional clustering method and maintains its accuracy.

• 한국데이터정보과학회:학술대회논문집
• /
• 2003.05a
• /
• pp.77-88
• /
• 2003

Cluster analysis has been widely used in many applications, such that data analysis, pattern recognition, image processing, etc. But clustering requires many hours to get clusters that we want, because it is more primitive, explorative and we make many data an object of cluster analysis. In this paper we propose a new clustering method, 'Clustering algorithm using a center of gravity for grid-based sample'. It is more fast than any traditional clustering method and maintains accuracy. It reduces running time by using grid-based sample and keeps accuracy by using representative point, a center of gravity.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.25 no.6
• /
• pp.36-41
• /
• 2002

We consider the Fuzzy clustering which is devised for partitioning a set of objects into a certain number of groups by assigning the membership probabilities to each object. The researches carried out in this field before show that the Fuzzy clustering concept is involved so much that for a certain set of data, the main purpose of the clustering cannot be attained as desired. Thus we propose a new objective function, named as Fuzzy-Entroppy Function in order to satisfy the main motivation of the clustering which is classifying the data clearly. Also we suggest Mean Field Annealing Algorithm as an optimization algorithm rather than the ISODATA used traditionally in this field since the objective function is changed. we show the Mean Field Annealing Algorithm works pretty well not only for the new objective function but also for the classical Fuzzy objective function by indicating that the local minimum problem resulted from the ISODATA can be improved.