固定二元值線性四元樹編碼法

Henry Ker-Chnag Chang, 張 智偉

Research output: Contribution to journalJournal Article peer-review

Abstract

     本文對於線性四元樹提出一新的線性編碼法,稱之為固定二元值線性四 元樹編碼法(Fixed Binary Linear Quadtree Coding Scheme,簡稱FBLQ),目的在改善 以往線性四元樹佔用記憶體過大的缺失。本編碼法設定時間複雜度在0(N),(N為 影像中的像元總數)的條件下,運用分支判斷法的技巧及摩頓數的方式完成。為 驗證本文所提之FBLQ方法的特性,本文以兩組不同解析度之空間資料進行實 驗,並與其他三種方法比較,實驗結果顯示,FBLQ方法確能比其他三個編碼法 所需的儲存空間要小;此外考慮空間資料最差情形下FBLQ所佔空間之理論值亦 為最小,故從實驗與理論分析的結果說明FBLQ方法的特性與優越性。
     In this paper, a new linear quadtree coding scheme. Fixed Binary Linear QuadtreeCoding (FBLQ), is proposed. The purpose is to improve the requirement of memory space.While the time complexity of FBLQ is limited by 0(N), N being the amount of imagepixels, the proposed FBLQ is developed on the bases of Morton sequence and the traversalorder of breadth first. A branch discrimination routine is provided to separate informationand indicator for nodes on a quadtree. Two raster images of different resolutions are testedto reveal the advantage of the proposed FBLQ scheme. The experiments arc compared withother three methods. Empirical tests demonstrate that the required memory space for theproposed FBLQ scheme is the least. Theoretical analysis for the worst case is also providedand compared. Results of both comparisons verify the feasibility of the proposed FBLQscheme. The proposed FBLQ scheme is not only suitable for binary imaghes but also usefulfor multicolor images.
Original languageChinese (Traditional)
Pages (from-to)67-77
Journal國立臺灣大學工程學刊
Issue number63
StatePublished - 1995

Cite this