精進寵物影像辨識效能工具開發計畫;Developing Tools for Enhancing Pet Image Recognition Performance

NCUIR > College of Electrical Engineering & Computer Science > Department of Computer Science and information Engineering > Research Project > Item 987654321/99046

Please use this identifier to cite or link to this item: https://ir.lib.ncu.edu.tw/handle/987654321/99046

Title:	精進寵物影像辨識效能工具開發計畫;Developing Tools for Enhancing Pet Image Recognition Performance
Authors:	陳慶瀚
Contributors:	國立中央大學資訊工程學系
Keywords:	寵物管理;身分識別;鼻紋;AI取像;Pet management;identity recognition;nose print;AI imaging
Date:	2026-01-22
Issue Date:	2026-01-23 16:30:31 (UTC+8)
Publisher:	農業部
Abstract:	開發一套穩定、易用且高效率的狗貓鼻紋取像和身分識別手機應用程式(APP)。以滿足：(1)顯著改善寵物鼻紋影像擷取過程中常見的模糊、解析度不足、角度偏差等問題，提升註冊和辨識之鼻紋影像品質。(2)建立一套鼻紋影像品質即時評估與拍攝引導機制，輔助使用者拍出符合標準的鼻紋照片。(3) 建立雲端資料庫管理系統，作為蒐集狗貓鼻紋影像資料庫以供日後AI模型訓練。本計畫預期將在以下六個面向產出具體可衡量之效益：1.學術成就發表國際期刊論文至少1篇、國內研討會論文1篇，深化動物AI視覺與農業管理領域之交叉應用。2.技術創新成就開發具備即時品質評估與拍攝引導功能的鼻紋影像擷取和上傳APP，有效提升拍攝品質並降低失敗率，克服現行手機拍攝鼻紋影像品質問題。提供一般使用者也能成功完成寵物鼻紋影像註冊與身分識別。3. 經濟效益推動非侵入式寵物辨識技術商轉，可衍生應用於保險理賠、寵物登記、醫療紀錄管理等領域，預估可創造潛在市場產值每年逾千萬元。有望吸引保險業、寵物醫療業、寵物用品產業等業者投入合作與投資，進一步擴大AI於農業及動物管理產業之應用版圖。4. 社會影響提升寵物登記覆蓋率與飼主責任意識，有助政府落實源頭管理政策與流浪犬減量目標，間接改善民生與公共安全。降低晶片植入爭議與成本，提升民眾接受度，支持「零撲殺政策」長期施行，符合動物保護與永續社會發展精神。5.非研究類成就培育1名博士研究人才和3名碩士研究人才，建立AI應用於動物辨識領域之跨域實作能力。促進政府與學研機構合作，建立可延展至家畜、實驗動物辨識系統的技術架構，支援政策工具革新。6.科技政策管理本技術可作為政策性動物登記制度之替代方案，提供農業主管機關導入新型登記制度的實證依據與技術工具。為政府日後制定或修訂《動物保護法》或寵物管理政策提供科技支撐與資料佐證。 ;To develop a stable, user-friendly, and highly efficient mobile application (APP) for dog and cat nose print image acquisition and identification. This APP aims to achieve the following:(1) Significantly improve the quality of nose print images for registration and identification by addressing common issues such as blur, insufficient resolution, and angular deviation during pet nose print image capture. (2) Establish a real-time nose print image quality assessment and shooting guidance mechanism to assist users in capturing standard-compliant nose print photos. (3) Build a cloud database management system to collect dog and cat nose print image data for future AI model training.This project is expected to yield measurable benefits in the following six areas:1. Academic AchievementsWe aim to publish at least one international journal paper and one domestic conference paper, deepening the cross-application of AI vision in the animal and agricultural management fields.2. Technological Innovation AchievementsWe will develop a nose print image acquisition and upload APP with real-time quality assessment and shooting guidance features. This will effectively improve shooting quality and reduce failure rates, overcoming current issues with mobile phone-captured nose print image quality. It will enable general users to successfully complete pet nose print image registration and identification.3. Economic BenefitsBy promoting the commercialization of non-invasive pet identification technology, we anticipate potential market output value exceeding ten million NTD annually. This technology can be derivatively applied to areas such as insurance claims, pet registration, and medical record management. We expect to attract collaboration and investment from the insurance, pet healthcare, and pet product industries, further expanding the application of AI in the agriculture and animal management sectors.4. Social ImpactThis initiative will increase pet registration coverage and pet owner responsibility awareness, helping the government implement source management policies and achieve stray dog reduction goals, indirectly improving public welfare and safety. It will also reduce disputes and costs associated with microchip implantation, enhance public acceptance, and support the long-term implementation of the "zero euthanasia policy," aligning with animal protection and sustainable social development principles.5. Non-Research AchievementsWe will cultivate one doctoral and three master's research talents, building cross-domain practical capabilities for AI applied in animal identification. We will also foster collaboration between government and academic research institutions to establish a technical architecture extendable to livestock and laboratory animal identification systems, supporting policy tool innovation.6. Technology Policy ManagementThis technology can serve as an alternative for policy-driven animal registration systems, providing agricultural authorities with empirical evidence and technical tools for introducing new registration systems. It will offer technological support and data validation for future government formulation or revision of the "Animal Protection Act" or pet management policies.
Relation:	財團法人國家實驗研究院科技政策研究與資訊中心
Appears in Collections:	[Department of Computer Science and information Engineering] Research Project

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	4	View/Open

社群 sharing

Loading...