计算机研究与发展
計算機研究與髮展
계산궤연구여발전
JOURNAL OF COMPUTER RESEARCH AND DEVELOPMENT
2015年
7期
1692-1704
,共13页
胡昭华%袁晓彤%李俊%何军
鬍昭華%袁曉彤%李俊%何軍
호소화%원효동%리준%하군
视觉跟踪%核稀疏%多特征联合%粒子滤波%重叠分块
視覺跟蹤%覈稀疏%多特徵聯閤%粒子濾波%重疊分塊
시각근종%핵희소%다특정연합%입자려파%중첩분괴
visual tracking%kernel sparse representation%multi-feature association%particle filter%overlapped fragment
大多数现有的基于稀疏表示的跟踪器仅采用单个目标特征来描述感兴趣的目标,因而在处理各种复杂视频时不可避免会出现跟踪不稳定的情况。针对这个问题,提出一种基于多特征联合稀疏表示的粒子滤波跟踪算法。该算法的主要思想是对随时间不断更新的字典模板和抽样粒子的局部块依据其位置进行分类,用字典中所有类别块对抽样粒子的局部块进行稀疏表示,而仅用与字典中具有相同类别的局部块及表示系数进行重构,根据重构误差构建似然函数以确定最佳粒子(候选目标),实现对目标的精确跟踪。该方法不仅实现了局部块的结构稀疏性,而且充分考虑了粒子之间的依赖关系,提高了跟踪精度。将算法进一步推广到采用基于核的多种特征描述,经混合范数约束并利用 KAPG (kernelizable accelerated proximal gradient )方法求解联合特征的稀疏系数。定性和定量的实验结果均表明该算法在目标发生遮挡、旋转、尺度变化、快速运动、光照变化等各种复杂情况下,依然可以准确地跟踪目标。
大多數現有的基于稀疏錶示的跟蹤器僅採用單箇目標特徵來描述感興趣的目標,因而在處理各種複雜視頻時不可避免會齣現跟蹤不穩定的情況。針對這箇問題,提齣一種基于多特徵聯閤稀疏錶示的粒子濾波跟蹤算法。該算法的主要思想是對隨時間不斷更新的字典模闆和抽樣粒子的跼部塊依據其位置進行分類,用字典中所有類彆塊對抽樣粒子的跼部塊進行稀疏錶示,而僅用與字典中具有相同類彆的跼部塊及錶示繫數進行重構,根據重構誤差構建似然函數以確定最佳粒子(候選目標),實現對目標的精確跟蹤。該方法不僅實現瞭跼部塊的結構稀疏性,而且充分攷慮瞭粒子之間的依賴關繫,提高瞭跟蹤精度。將算法進一步推廣到採用基于覈的多種特徵描述,經混閤範數約束併利用 KAPG (kernelizable accelerated proximal gradient )方法求解聯閤特徵的稀疏繫數。定性和定量的實驗結果均錶明該算法在目標髮生遮擋、鏇轉、呎度變化、快速運動、光照變化等各種複雜情況下,依然可以準確地跟蹤目標。
대다수현유적기우희소표시적근종기부채용단개목표특정래묘술감흥취적목표,인이재처리각충복잡시빈시불가피면회출현근종불은정적정황。침대저개문제,제출일충기우다특정연합희소표시적입자려파근종산법。해산법적주요사상시대수시간불단경신적자전모판화추양입자적국부괴의거기위치진행분류,용자전중소유유별괴대추양입자적국부괴진행희소표시,이부용여자전중구유상동유별적국부괴급표시계수진행중구,근거중구오차구건사연함수이학정최가입자(후선목표),실현대목표적정학근종。해방법불부실현료국부괴적결구희소성,이차충분고필료입자지간적의뢰관계,제고료근종정도。장산법진일보추엄도채용기우핵적다충특정묘술,경혼합범수약속병이용 KAPG (kernelizable accelerated proximal gradient )방법구해연합특정적희소계수。정성화정량적실험결과균표명해산법재목표발생차당、선전、척도변화、쾌속운동、광조변화등각충복잡정황하,의연가이준학지근종목표。
Most existing sparse representation based trackers only use a single feature to describe the objects of interest and tend to be unstable when processing challenging videos .To address this issue , we propose a particle filter tracker based on multiple feature joint sparse representation .The main idea of our algorithm is to partition each particle region into multiple overlapped image fragments . Eevery local fragment of candidates is sparsely represented as a linear combination of all the atoms of dictionary template that is updated dynamically and is merely reconstructed by the local fragments of dictionary template located at the same position . The weights of particles are determined by their reconstruction errors to realize the particle filter tracking .Our method simultaneously enforces the structural sparsity and considers the interactions among particles by using mixed norms regularization . We further extend the sparse representation module of our tracker to a multiple kernel joint sparse representation module which is efficiently solved by using a kernelizable accelerated proximal gradient (KAPG ) method . Both qualitative and quantitative evaluations demonstrate that the proposed algorithm is competitive to the state‐of‐the‐art trackers on challenging benchmark video sequences with occlusion ,rotation ,shifting and illumination changes .