• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.55-58
• /
• 2024

In this paper, we propose a design and implementation of big data-based fine dust anomaly detection machine learning system. The proposed is system that classifies the fine dust air quality index through meteorological information composed of fine dust and big data. This system classifies fine dust through the design of an anomaly detection algorithm according to the outliers for each air quality index classification categories based on machine learning. Depth data of the image collected from the camera collects images according to the level of fine dust, and then creates a fine dust visibility mask. And, with a learning-based fingerprinting technique through a mono depth estimation algorithm, the fine dust level is derived by inferring the visibility distance of fine dust collected from the monoscope camera. For experimentation and analysis of this method, after creating learning data by matching the fine dust level data and CCTV image data by region and time, a model is created and tested in a real environment.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.15 no.3
• /
• pp.35-41
• /
• 2015

In order to solve the problems that fingerprinting technologies and image identification technologies for digital music and videos identification scanned have not yet been enough to identify scanned cartoon books, the copyright project granted from the Korea Copyright Commission(Project title: Identification and Copy Protection Technology of Scanned Text/Comic Books) secured basic solutions for scanned cartoon books. The first result of the research project in 2013 was focused on the identification technology applied to abnormally scanned cartoon books. As subsequent research for the project, studies in the year of 2014 has been to provide solutions to identify cartoon books which abnormally scanned and occurred with transformations and distortions and solutions to protect the system from trials to devitalize functions of the identification system through various attacks. This paper is to propose a model for a system for identifying abnormally scanned cartoon books as a way of enhancing the identification rates for identifying abnormal bookscans and transformation attacks.

• Journal of Embryo Transfer
• /
• v.16 no.3
• /
• pp.183-191
• /
• 2001

This study was designed to establish the superior method for IVF embryos from DNA marker-proved Hanwoo cattle. DNA markers related to marbling score were identified using DNA fingerprinting with Ml3 probe and restriction enzyme Hae III. Oocytes were aspirated from unstimulated. immature ovarian follicles using a combined method of rectal ovarian-palpation and transvaginal ultrasound-guidance(6.5MHz) under local abesthesia. The aspirated oocytes were washed twice with fresh D-PBS containing 5% FBS and were rewashed 4 to 5 times with TCM-199 containing 5% FBS. A morphological grade of I to IV was assigned to each oocyte. Data were analysed using the GLM procedure of SAS. Mean number of follicles identified on ultrasound was 5.5 $\pm$2.9 in right and 4.3 $\pm$2.8 in left ovaries, respectively. The highest follicles(16.6$\pm$2.6) were found in 5101 cow compared to others. Recovery rate of follicular oocytes in individual cow was highest in 5101 cow with 89.3% in > 2mm and 94.0% in $\leq$ 2mm follicles. Total recovery rate was significantly(P<0.01) higher in $\leq$ 2mm(85.7%, 130/154) than > 2mm follicles(74.2%, 201/271). Significantly more oocytcs of Grade IV were recovered from > 2mm follicles. Mean number follicles recovered was 4.8$\pm$3.7. 3.0$\pm$3.4 and 0.3$\pm$0.6 in $\leq$2mm, 2~6mm and $\geq$6mm follicles. respectively. Our results imply that the more fertilizablc oocytes can be recovered from invisible-immature follicles by the combination of simultaneous rectal ovarian-palpation and ultrasound-guided approach in Hanwoo cattle.