- 學位論文
智能影像分析架構於空間域和壓縮域之應用
Application of Intelligent Image Analysis Framework in Spatial Domain and Compressed Domain
摘要
本研究的目標是發展一套可以使用在空間域和壓縮域圖像的智能影像分析架構，我們提供兩個成功案例來證明這個分析架構可行： “自我護理系統於術後傷口診斷的應用” 以及 “在H.265 / HEVC串流中使用警戒線進行移動物體的計數”。 第一個案例是用於開發一種對患者友好的自我護理系統，它可以幫助患者和醫療專業人員在沒有任何特殊醫療設備的情況下診斷術後傷口的狀態。使用該方法，超過90％的狀態評估結果是正確的，並且有超過80％的症狀評估結果與實際診斷一致。第二個案例可以在不完全對H.265 / HEVC串流解碼的前提下估算通過特定區域的移動物體數量。 兩種案例分別在空間域和壓縮域中識別模式。第一個案例顯示通過分析空間域中的影像該系統可以在實際的自我護理條件下表現良好。第二個案例則顯示可以通過分析壓縮域中的影像來快速偵測並計算移動物體的數量。
並列摘要
The objective of this research is to develop an intelligent image framework which can be used in spatial domain and compressed domain. There are two successful case studies to prove that our approach works: Surgical Wounds Assessment System for Self-Care and Moving Object Counting Using a Tripwire in H.265/HEVC Bitstreams for Video Surveillance. The first case is used to develop a patient-friendly self-care system which can help both patients and medical professionals to ensure the state of the surgical wounds without any special medical equipment. With the proposed method, more than 90% state assessment results are correct and more than 80% symptom assessment results consistent with the actual diagnosis. The second case, is used to estimate the number of moving objects that passes through a specific area without fully decoding the H.265/HEVC bitstreams. The two cases respectively recognize pattern in the spatial domain and in the compressed domain. The first case shows that this system could perform well in the practical self-care scenario by analyzing image in the spatial domain. The second research results show that it can fast detect and count moving object by analyzing image in the compressed domain.