Myoung-Beom Chung

Intelligent copyright protection system using a matching video retrieval algorithm

Many video service sites headed by YouTube know what content requires copyright protection. However, they lack a copyright protection system that automatically distinguishes whether uploaded videos contain legal or illegal content. Existing protection techniques use content-based retrieval methods that compare the features of video. However, if the video encoding has changed in resolution, bit-rate or codec, these techniques do not perform well. Thus, this paper proposes a novel video matching algorithm even if the type of encoding has changed. We also suggest an intelligent copyright protection system using the proposed algorithm. This can serve to automatically prevent the uploading of illegal content. The proposed method has represented the accuracy of 97% with searching algorithm in video-matching experiments and 98.62% with automation algorithm in copyright-protection experiments. Therefore, this system could form a core technology that identifies illegal content and automatically excludes access to illegal content by many video service sites.

An algorithm that minimizes audio fingerprints using the difference of Gaussians

Recently, many audio search sites headed by Google have used audio fingerprinting technology to search for the same audio and protect the music copyright using one part of the audio data. However, if there are fingerprints per audio file, then the amount of query data for the audio search increases. In this paper, we propose a novel method that can reduce the number of fingerprints while providing a level of performance similar to that of existing methods. The proposed method uses the difference of Gaussians which is often used in feature extraction during image signal processing. In the experiment, we use the proposed method and dynamic time warping and undertake an experimental search for the same audio with a success rate of 90%. The proposed method, therefore, can be used for an effective audio search.

Identical-video retrieval using the low-peak feature of a video’s audio information

The recognition and retrieval of identical videos by combing through entire video files requires a great deal of time and memory space. Therefore, most current video-matching methods analyze only a part of each video’s image frame information. All these methods, however, share the critical problem of erroneously categorizing identical videos as different if they have merely been altered in resolution or converted with a different codec. This paper deals instead with an identical-video-retrieval method using the low-peak feature of audio data. The low-peak feature remains relatively stable even with changes in bit-rate or codec. The proposed method showed a search success rate of 93.7% in a video matching experiment. This approach could provide a technique for recognizing identical content on video file share sites.

Data-Sharing Method for Multi-Smart Devices at Close Range

We proposed a useful data-sharing method among multi-smart devices at close range using inaudible frequencies and Wi-Fi. The existing near data-sharing methods mostly use Bluetooth technology, but these methods have the problem of being unable to be operated using different operating systems. To correct this flaw, the proposed method that uses inaudible frequencies through the inner speaker and microphone of smart device can solve the problems of the existing methods. Using the proposed method, the sending device generates trigger signals composed of inaudible sound. Moreover, smart devices that receive the signals obtain the shared data from the sending device through Wi-Fi. To evaluate the efficacy of the proposed method, we developed a near data-sharing application based on the trigger signals and conducted a performance evaluation experiment. The success rate of the proposed method was 98.8%. Furthermore, we tested the user usability of the Bump application and the proposed method and found that the proposed method is more useful than Bump. Therefore, the proposed method is an effective approach for sharing data practically among multi-smart devices at close range.

Detection of a Robust High-Frequency Range via Noise Analysis in a Real-World Environment

Recently, many studies have been conducted on using inaudible high frequencies for wireless communication based on smart devices and data transmission algorithms. However, many studies have identified a problem such that transmission accuracy is extremely low because of ambient noise in the real-life environment. To solve this problem, we proposed an application and server system. The proposed application can gather many sounds, including those with high frequencies; the gathered high frequencies are sent to a server system that can detect a robust high-frequency range via statistical processing. We tested the proposed application’s ability to gather noise and high frequencies for a certain period of time to evaluate performance. According to the testing results, the proposed application and server system could detect a robust high-frequency range via noise analysis in real life. Therefore, the proposed application and server could be a useful technology for future research on inaudible high frequencies.

Mobile Book Management Application Using Camera of Smart Phone

This thesis suggests an application that will allow users to manage their books easily by using an open API based book information available from smart phones or on-line. A camera of a smart phone can recognize bar-codes and QR-codes on books automatically and help users retrieve more detailed information through the open API based on thus recognized information. The suggested application will help users save the provided content of a book in Scan List or My List, and create a memo for the page users read last, and write a simple comment when they finished the book. In other words, the suggested application will provide useful functions for users to manage books effectively.

Data communication method based on inaudible sound at near field

In this paper, we proposed data communication method based on inaudible sound which the most people cannot listen to at near field. The proposed method can transmit some bits data as using mixed high frequencies which are mapped at each data bit. Besides, the proposed method increases the accuracy of data transmission as using cyclic redundancy check for prevention of transmission error. To evaluate the performance of the proposed method, we tested with the proposed method and the result showed 97.5% accuracy. Therefore, the proposed method could be applied at various wireless data communication technology fields.