국제협력사업
| 연구번호 | RS-2023-00269693 | ||
|---|---|---|---|
| 연구사업명 | Local demonstration and marketization technology for AI-based screening and control systems suitable for the ASEAN region | ||
| 연구기간 | 2023.06.30 ~ 2027.12.31 | 연구책임자 | - |
| 사업예산 | 2,560 백만원 | 대상국 | Indonesia |
| 기술분류 | Urban space | 영상자료 | https://www.markany.com/aivideoanalyticshttps://www.youtube.com/watch?v=XUSwv-YPXdwhttps://www.youtube.com/watch?v=S8wMMZyna_4&t=28s |
| 연구보고서 | 기술개발홍보자료 | ||
| 연구내용 |
Development and demonstration of an intelligent screening control system to control abnormal behavior (5-types) suitable for ASEAN
○ Building raw data for screening - Preparation of scenarios for each type of abnormal behavior and proceeding with actor selection - CCTV raw data collection in KOTA Command Center - Set the data construction definition and scale, and perform the acquisition data scale and basic design for each class ○ Data purification and processing - Build source data by refining raw data (removing duplicate files, de-identifying, cropping, removing duplicate data, etc.) - Perform data cutting and de-identification tasks based on semantic accuracy, diversity, and de-identification of important information from raw data - Raw data for detection classes is created as a source dataset after processing personal information without identifying personal information through personal information de-identification processing - Use work tools to perform labeling tasks in accordance with labeling guidelines - After checking the labeling, classify classes for the appropriate data model and store the labeling data ○ Building a learning model to improve object recognition technology - Prior research and candidate group setting for learning models - Establishing learning models related to general (normal) behavior - Building a learning model for abnormal behavior (loitering, intrusion, fire, motorcycle theft, and gangsters) - Conducting joint research and collaboration with Bandung Institute of Technology according to local conditions - Selection of quality indicators for each candidate group of learning models - Verification of learning models and application of solutions based on quality indicators |
||
연구기관명
| 번호 | 연구기관명 | 구분 | 전화번호 | |
|---|---|---|---|---|
| 1 | Markany | 주관연구기관 | ||
| 2 | ITB | 해외협력기관 | ||
| 3 | UNJ | 해외협력기관 |