本文選題：目標(biāo)跟蹤 + 卷積特征��； 參考：《中國圖象圖形學(xué)報》2017年09期

【摘要】：目的針對現(xiàn)實場景中跟蹤目標(biāo)的快速運動、旋轉(zhuǎn)、尺度變化、遮擋等問題,提出了基于卷積特征的核相關(guān)自適應(yīng)目標(biāo)跟蹤的方法。方法利用卷積神經(jīng)網(wǎng)絡(luò)提取高、低層卷積特征并結(jié)合本文提出的核相關(guān)濾波算法計算并獲得高底兩層卷積特征響應(yīng)圖。采用Coarse-to-Fine方法對目標(biāo)位置進行估計,在學(xué)習(xí)得到1維尺度核相關(guān)濾波器估計尺度的基礎(chǔ)上實時更新高低兩層核相關(guān)濾波器參數(shù),以實現(xiàn)自適應(yīng)的目標(biāo)跟蹤。結(jié)果實驗選取公開數(shù)據(jù)集中的典型視頻序列進行跟蹤,測試了算法在目標(biāo)尺度發(fā)生變化、遮擋、旋轉(zhuǎn)等復(fù)雜場景下的跟蹤性能并與多種優(yōu)秀的跟蹤算法在平均中心誤差、平均重疊率等指標(biāo)上進行了定量比較,在Singer1、Car4、Jogging、Girl、Football以及MotorRolling視頻圖像序列上的中心誤差分別為8.71、6.83、3.96、3.91、4.83、9.23,跟蹤重疊率分別為0.969、1.00、0.967、0.994、0.967、0.512。實驗結(jié)果表明,本文算法與原始核相關(guān)濾波算法相比,平均中心位置誤差降低20%,平均重疊率提高12%。結(jié)論采用卷積神經(jīng)網(wǎng)絡(luò)提取高低兩層卷積特征,高層卷積特征用于判別目標(biāo)和背景,低層卷積特征用于預(yù)測目標(biāo)位置并通過Coarse-to-Fine方法對目標(biāo)位置進行精確的定位,較好地解決了由于目標(biāo)的旋轉(zhuǎn)和尺度變化帶來的跟蹤誤差大的問題,提高了跟蹤性能并能夠?qū)崟r更新學(xué)習(xí)。在目標(biāo)尺度發(fā)生變化、遮擋、光照條件改變、目標(biāo)快速運動等復(fù)雜場景下仍表現(xiàn)出較強的魯棒性和適應(yīng)性。
[Abstract]:Aim to solve the problems of fast moving, rotation, scale change, occlusion and so on, a kernel correlation adaptive target tracking method based on convolution feature is proposed. Methods the high and low level convolution features were extracted by convolution neural network and the response map of the high bottom two-layer convolution feature was obtained by combining the kernel correlation filtering algorithm proposed in this paper. The Coarse-to-Fine method is used to estimate the location of the target. Based on the learning of the estimation scale of the 1-dimensional kernel correlation filter, the parameters of the high and low two layers correlation filter are updated in real time to achieve adaptive target tracking. Results the typical video sequences in the open data set were selected for tracking, and the tracking performance of the algorithm in complex scenes such as target scale change, occlusion, rotation and so on was tested, and the average center error of the algorithm was compared with that of many excellent tracking algorithms. Quantitative comparison was made on the average overlap rate. The central errors in Singer 1 / Car4 Jogging GirlsFootball and MotorRolling video sequences were 8.71 / 6.83 / 3.963.91 / 4.839.23, respectively, and the tracking overlap rates were 0.9691.000.9670.9940.967/ 0.967/ 0.512, respectively. The experimental results show that compared with the original kernel correlation filtering algorithm, the average center position error is reduced by 20% and the average overlap rate is increased by 12%. Conclusion the convolution neural network is used to extract the convolution feature of high and low layers, the high level convolution feature is used to distinguish the target and the background, and the lower level convolution feature is used to predict the target position and to locate the target position accurately by Coarse-to-Fine method. The problem of large tracking error caused by target rotation and scale change is well solved, and the tracking performance is improved and the learning can be updated in real time. It still shows strong robustness and adaptability in complex scenes such as the change of target scale, occlusion, light condition change, fast moving of target and so on.
【作者單位】： 中北大學(xué)計算機與控制工程學(xué)院;
【基金】：山西省自然科學(xué)基金項目(2013011017-6)~~
【分類號】：TP391.41

【相似文獻】

相關(guān)期刊論文 前10條

1 H.Meerkamm,St.Sander,R.-D.Kasan,陳巍;一個建立特征圖書庫的方案[J];工程設(shè)計學(xué)報;2001年04期

2 陳佳;;基于眼睛特征圖的眼睛檢測系統(tǒng)的設(shè)計與實現(xiàn)[J];南陽師范學(xué)院學(xué)報;2008年06期

3 魏永偉;曹娟;汪國昭;;平面三次混合雙曲多項式曲線的特征圖判別[J];計算機輔助設(shè)計與圖形學(xué)學(xué)報;2010年05期

4 孫正興，，丁秋林;基于特征的CAD／CAM集成技術(shù)評述與研究[J];計算機應(yīng)用;1994年06期

5 祝國旺;鐘毅芳;周濟;余俊;;零件形狀特征圖樹分解原理及應(yīng)用[J];交通與計算機;1993年02期

6 段冠新;;在SolidWorks中庫特征的創(chuàng)建和使用[J];機械工人.冷加工;2006年02期

7 葛江華,隋秀凜,鄒世浩;基于特征的電站閥門CAD/CAPP/CAM集成系統(tǒng)[J];哈爾濱理工大學(xué)學(xué)報;2002年02期

8 李凈凈;張基宏;梁永生;;視覺注意模型中權(quán)值估計與特征整合[J];深圳信息職業(yè)技術(shù)學(xué)院學(xué)報;2012年03期

9 徐志;關(guān)宏志;嚴海;陳二慧;;基于圖像顯著性特征的交通標(biāo)志注視點預(yù)測方法[J];武漢理工大學(xué)學(xué)報(交通科學(xué)與工程版);2014年01期

10 許本勝;黃美發(fā);蘇慶勇;王燦;;基于特征幾何變動的三維公差綜合建模方法研究[J];制造業(yè)自動化;2014年06期

相關(guān)碩士學(xué)位論文 前3條

1 楊凡;基于層次特征的視覺注意模型研究[D];華中科技大學(xué);2016年

2 魏艾;基于全局和局部深度特征的圖像重排序方法研究[D];中國科學(xué)技術(shù)大學(xué);2017年

3 陳超;基于個人微博特征的Timeline生成研究[D];內(nèi)蒙古科技大學(xué);2015年

本文編號：1951885