









摘" 要: 隨著視頻應用和新興業務的快速發展,對視頻編碼速度和質量的要求也不斷提高。為了降低H.265/HEVC的幀內編碼復雜度,提出一種基于最有可能模式(MPM)的模糊搜索算法,通過減少搜索候選模式的數量來降低計算復雜度;同時提出一種簡化編碼單元劃分過程的方法,利用相鄰編碼單元率失真代價計算的閾值,提前終止編碼單元劃分,避免了傳統算法的遍歷劃分,提高了編碼效率。實驗結果表明,所提算法與HEVC傳統模型比較,能夠平均降低39.38%的編碼時間,而碼率只增加了1.62%,峰值信噪比差值僅降低0.085 dB。在保證視頻質量的前提下,所提算法大幅降低了編碼的復雜度。
關鍵詞: HEVC; 視頻編碼; 幀內編碼; 最有可能模式; 編碼單元劃分; 率失真優化
中圖分類號: TN919.81?34" " " " " " " " " " " " "文獻標識碼: A" " " " " " " " " " " "文章編號: 1004?373X(2025)08?0051?05
Research on fast intraframe coding based on H.265/HEVC
MA Zhenhua1, JIA Huayu1, LUO Biao2
(1. College of Electrical and Power Engineering, Taiyuan University of Technology, Taiyuan 030024, China; 2. Accelink Technologies Co., Ltd., Wuhan 430074, China)
Abstract: With the rapid development of video applications and quality businesses, the requirements for video encoding speed and quality are increasing constantly. In order to reduce the H.265/HEVC (high efficiency video coding) complexity of intra frame coding, a fuzzy search algorithm based on the most probable mode (MPM) is proposed, which can reduce computational complexity by reducing the number of candidate search patterns. At the same time, a method for simplifying the process of encoding unit division is proposed, which can use the threshold calculated by the rate distortion cost of adjacent encoding units to terminate encoding unit division in advance, avoiding the traditional algorithm's traversal division and improving encoding efficiency. The experimental results show that compared with the traditional HEVC model, the proposed algorithm can reduce the encoding time by an average of 39.38%, while the bit rate is only increased by 1.62%, and the peak signal?to?noise ratio difference is only reduced by 0.085 dB. Under the premise of ensuring video quality, the proposed algorithm can greatly reduce the encoding complexity.
Keywords: HEVC; video coding; intra coding; most probable mode; coding unit partition; rate?distortion optimization
0" 引" 言
隨著人們對高清視頻的需求不斷增大,視頻編碼聯合協助小組推出了新一代的高效視頻編碼(HEVC)標準[1],也稱H.265,通過引入四叉樹結構、更多預測模式等新算法,并新增編碼單元(Coding Unit, CU)、預測單元(Prediction Unit, PU)和變換單元(Transform Unit, TU)等概念,使HEVC能夠更有效地處理視頻數據,實現更高質量的視頻壓縮,提高了50%以上的率失真性能。但這種性能改進是以增加編碼復雜度為代價的,為了滿足當前市場對實時高清視頻的需求,降低HEVC的編碼復雜度已成為當前研究的熱點問題。近年來,許多學者致力于降低HEVC幀內編碼復雜度的研究,主要集中在三類方法:基于預測模式選擇、基于編碼單元劃分、基于機器學習的方法。
在第一類方法中,研究者們嘗試通過不同的途徑來優化預測模式的選擇,減少編碼復雜度。……