• Smart Media Journal
• /
• v.12 no.9
• /
• pp.159-170
• /
• 2023

It takes a lot of time and manpower to search for the missing. As part of the solution, a missing person search AI system was implemented using a YOLO-based model. In order to train object detection models, the model was learned by collecting recognition images (road fixation) of drone mobile objects from AI-Hub. Additional mountainous terrain datasets were also collected to evaluate performance in training datasets and other environments. In order to optimize the missing person search AI system, performance evaluation based on model size and hyperparameters and additional performance evaluation for concerns about overfitting were conducted. As a result of performance evaluation, it was confirmed that the YOLOv5-L model showed excellent performance, and the performance of the model was further improved by applying data augmentation techniques. Since then, the web service has been applied with the YOLOv5-L model that applies data augmentation techniques to increase the efficiency of searching for missing people.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.2
• /
• pp.775-792
• /
• 2015

Person re-identification is an important and challenging task in computer vision with numerous real world applications. Despite significant progress has been made in the past few years, person re-identification remains an unsolved problem. This paper presents a novel appearance-based approach to person re-identification. The approach exploits region covariance matrix and color histograms to capture the statistical properties and chromatic information of each object. Robustness against low resolution, viewpoint changes and pose variations is achieved by a novel signature, that is, the combination of Log Covariance Matrix feature and HSV histogram (LCMH). In order to further improve re-identification performance, third-party image sets are utilized as a common reference to sufficiently represent any image set with the same type. Distinctive and reliable features for a given image set are extracted through decision boundary between the specific set and a third-party image set supervised by max-margin criteria. This method enables the usage of an existing dataset to represent new image data without time-consuming data collection and annotation. Comparisons with state-of-the-art methods carried out on benchmark datasets demonstrate promising performance of our method.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.1
• /
• pp.81-88
• /
• 2011

Augmented multimedia is a technology to provide additional informations to mobile devices when multimedia contents like video, audio and images are being played. Person Wide Web, PWW, is a scheme for acquiring a link and browsing a corresponding web pages on mobile devices, in which the link is attached any object and space in real world. In this paper we proposed an augmented multimedia E-commerce application system based on PWW scheme which can browse additional informations from video play on public spaces, and implemented on Microsoft Silverlight platform. We showed that the proposed system can support effectively the augmented multimedia E-commerce.

• Journal of the Korea Computer Graphics Society
• /
• v.23 no.3
• /
• pp.9-17
• /
• 2017

This study proposes a third-person virtual reality content production pipeline to provide users with a new experience and presence in a new virtual reality environment. For this purpose, we first create third-person virtual reality content, which includes a story, fun factors, and game characteristic. It consists of a tutorial scene in which a user can pre-learn the proposed interface suitable for a third person different from existing virtual reality content and a content scene that achieves its purpose by using game factors based on the background story. Next, we design an interface suitable for the third-person virtual reality content. This study proposes an interface in which users can interact with a virtual environment or object by using their hand. The proposed interface consists of three steps: character movement, virtual object selection with multiple selection, and 3D menu control using virtual space. Finally, through the survey experiment, third-person virtual reality content produced based on the proposed interface was confirmed to be easily controlled while ensuring high satisfaction.

• 제어로봇시스템학회:학술대회논문집
• /
• 1998.10a
• /
• pp.70-74
• /
• 1998

This paper describes a method for recognition of two-dimensional position of an object by use of aerial ultra-sonic sensor and signal processing technique, which would become a help for blind person or self-mobile robot. First, we have developed a method for measuring the time difference between the transmitted and the received burst wave by use of one ultrasonic transmitter and three receivers. Secondly, a new method is developed for measuring the distance to an object by use of M-sequence correlation method. Thirdly, a measurement method to obtain the position of an object is described by use of phase-arrayed ultrasonic sensor, which gives us a wide-range position determination in a short time.

• Asia Pacific Journal of Corpus Research
• /
• v.1 no.2
• /
• pp.15-26
• /
• 2020

This corpus-based study investigated third language acquisition by Taiwanese college students in learning Spanish clitic pronouns at beginning and intermediate levels. It examined the acquisition sequences of Spanish clitic pronouns of the Chinese-speaking learners whose second language was English and third language was Spanish. The results indicated that indirect object pronouns (OP) preceded direct OP (case), first person preceded third person OP (person), masculine preceded feminine OP (gender), and animate preceded inanimate OP (animacy). The findings presented similar patterns as those of previous studies on English-speaking learners of Spanish. In further comparisons of the target forms in Chinese, English, and Spanish, the results suggested that L1 Chinese had strong influence on L3 Spanish, which accounts for the challenges that Taiwanese learners of Spanish face as they learn the Spanish clitic pronouns in the beginning stage.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.11a
• /
• pp.814-817
• /
• 2020

The number of single-person households is increasing every year, and there are also high concerns about the crime and safety of single-person households. In particular, crimes targeting women are increasing. Although home surveillance camera applications, which are mostly used by single-person households, only provide intrusion detection functions, this service utilizes AI image recognition technologies such as face recognition and object detection to provide theft, violence, stranger and intrusion detection. Users can receive security-related notifications, relieve their anxiety, and prevent crimes through this service.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.17 no.2
• /
• pp.141-146
• /
• 2017

New infectious diseases such as MERS have been in need of many measures such as initial discovery, isolation, and crisis response. In addition, the culture of hospitals is changing, such as the general public 's visiting and Nursing Care Integration Services. However, as the qualifications and regulations of medical personnel in hospitals become rigid, overseas such as linens, wastes movements are replacing possible works with robots. we have developed a hospital logistics robot that can carry out various goods delivery within a hospital, and can move various kinds of objects safely to a desired location. In this thesis, we have studied a hospital logistics robot that can carry out various kinds of goods delivery within the hospital, and can move various kinds of objects such as waste, and linen safely to a desired location. The movement of a robot in a hospital may cause a collision between a person and an object, so that the collision must be prevented. In order to prevent collision, it is necessary to recognize whether or not an object exists in the movement path of the robot. And if there is an object, it should recognize whether it moves or not. In order to recognize human beings and objects, we recognize the person with face/body recognition technology and generate the context awareness of the object using 3D Vision image segmentation technology. We use the generated information to create a map that considers objects and person in the robot moving range. Thus, the robot can be operated safely and efficiently.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.7
• /
• pp.866-872
• /
• 2020

Indoor location-based services provide a service based on the distance between an object and a person. Recently, indoor location-based services are often implemented using inexpensive depth sensors such as Kinect. The depth sensor provides a function to measure the position of a person, but the position of an object must be acquired manually using a tool. To acquire a 3D position of an object, it requires 3D interaction, which is difficult to a general user. GUI(Graphical User Interface) is relatively easy to a general user but it is hard to gather a 3D position. This study proposes the Interactive LOcation Context Authoring Toolkit(ILOCAT), which enables a general user to easily acquire a 3D position of an object in real space using GUI. This paper describes the interaction design and implementation of ILOCAT.

• Convergence Security Journal
• /
• v.22 no.5
• /
• pp.11-20
• /
• 2022

As the Internet and web technology develop around mobile devices, image data contains various types of sensitive information such as people, text, and space. In addition to these characteristics, as the use of SNS increases, the amount of damage caused by exposure and abuse of personal information online is increasing. However, research on de-identification technology based on multi-type object detection for personal information protection is insufficient. Therefore, this paper proposes an artificial intelligence model that detects and de-identifies multiple types of objects using existing single-type object detection models in parallel. Through cutmix, an image in which person and text objects exist together are created and composed of training data, and detection and de-identification of objects with different characteristics of person and text was performed. The proposed model achieves a precision of 0.724 and mAP@.5 of 0.745 when two objects are present at the same time. In addition, after de-identification, mAP@.5 was 0.224 for all objects, showing a decrease of 0.4 or more.