目标跟踪中辅助目标的选择、跟踪与更新

刘畅; 赵巍; 刘鹏; 唐降龙

doi:10.16383/j.aas.2017.c160532

目标跟踪中辅助目标的选择、跟踪与更新

doi: 10.16383/j.aas.2017.c160532

刘畅^1,2,,
赵巍^1,2,,
刘鹏^1,2, ,,
唐降龙^1,2,

1.
哈尔滨工业大学计算机科学与技术学院哈尔滨 150001
2.
哈尔滨工业大学模式识别与智能系统研究中心哈尔滨 150001

基金项目:

国家自然科学基金 61672190

国家自然科学基金 61671175

详细信息

作者简介:
刘畅  哈尔滨工业大学计算机科学与技术学院博士研究生.主要研究方向为计算机视觉与模式识别.E-mail:magicallc@126.com

赵巍  哈尔滨工业大学计算机科学与技术学院副教授.2006年获得哈尔滨工业大学计算机应用技术博士学位.主要研究方向为计算机视觉与模式识别.E-mail:zhaowei@hit.edu.cn

唐降龙  哈尔滨工业大学计算机科学与技术学院教授.1995年获得哈尔滨工业大学计算机应用技术博士学位.主要研究方向为计算机视觉与模式识别.E-mail:tangxl@hit.edu.cn

通讯作者:
刘鹏哈尔滨工业大学计算机科学与技术学院副教授.2007年获得哈尔滨工业大学微电子与固体电子学博士学位.主要研究方向为计算机视觉与模式识别.本文通信作者.E-mail:pengliu@hit.edu.cn

计量
- 文章访问数: 2766
- HTML全文浏览量: 273
- PDF下载量: 1243
- 被引次数: 0
出版历程
- 收稿日期: 2016-07-18
- 录用日期: 2017-06-07
- 刊出日期: 2018-07-20

Auxiliary Objects Selecting, Tracking and Updating in Target Tracking

LIU Chang^{1,2
,},
ZHAO Wei^{1,2
,},
LIU Peng^{1,2
, ,},
TANG Xiang-Long^{1,2
,}

1.
College of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001
2.
Research Center of Pattern Recognition and Intelligence System, Harbin Institute of Technology, Harbin 150001

Funds:

National Natural Science Foundation of China 61672190

National Natural Science Foundation of China 61671175

More Information

Author Bio:
Ph. D. candidate at the College of Computer Science and Technology, Harbin Institute of Technology. His research interest covers computer vision and pattern recognition

Associate professor at the College of Computer Science and Technology, Harbin Institute of Technology. She received her Ph. D. degree in computer application technology from Harbin Institute of Technology in 2006. Her research interest covers computer vision and pattern recognition

Professor at the College of Computer Science and Technology, Harbin Institute of Technology. He received his Ph. D. degree in computer application technology from Harbin Institute of Technology in 1995. His research interest covers computer vision and pattern recognition

Corresponding author: LIU Peng Associate professor at the College of Computer Science and Technology, Harbin Institute of Technology. He received his Ph. D. degree in microelectronics and solid state electronics from Harbin Institute of Technology in 2007. His research interest covers computer vision and pattern recognition. Corresponding author of this paper

摘要

摘要: 在目标的动态背景中存在有利于目标跟踪的信息.使用辅助目标来描述目标的动态背景，在跟踪目标的同时建立辅助目标与目标之间的运动依赖关系.用多个辅助目标预测目标的位置，将辅助目标预测结果与目标跟踪器预测结果融合得到目标位置.方法包括，利用辅助目标与目标之间的运动依赖关系和辅助目标自身跟踪精度的辅助目标选择方法；使用多个低精度辅助目标跟踪器获得良好的目标预测结果的辅助目标跟踪和目标预测方法；以及根据目标跟踪结果更新辅助目标跟踪参数的方法.辅助目标具有多样性和独立性.基于辅助目标的目标跟踪方法可以与其他目标跟踪器结合，具有推广泛化能力.实验结果表明，辅助目标在目标跟踪中发挥重要作用，与其他目标跟踪方法对比结果表明，有更好的鲁棒性和目标跟踪精度.
- 目标跟踪 /
- 动态背景 /
- 辅助目标 /
- 跟踪器融合
Abstract: The dynamic background of images contains effective information that is favorable for object tracking. This paper uses auxiliary objects describing the target background. Motion dependencies between auxiliary objects and target itself are established first. Then the tracking of auxiliary objects by trackers with relative lower precisions predicts the target position. The position results predicted by auxiliary objects and by the basic tracker are fused to get a better result. Meanwhile, parameters updating is done according to the tracking result. Extensive experimental results show that auxiliary objects are effective for tracking the target. Comparison with other trackers proves the method's better robustness and precision.
- Object tracking /
- dynamic background /
- auxiliary objects /
- trackers fusion
注释:

1) 本文责任编委赖剑煌

HTML全文

图 1 辅助目标在目标跟踪中的作用示意图

Fig. 1 The effect of auxiliary objects in target tracking

下载: 全尺寸图片幻灯片

图 2 用辅助目标跟踪的目标跟踪方法的框架

Fig. 2 The framework of the target tracking with auxiliary objects

下载: 全尺寸图片幻灯片

图 3 候选辅助目标选择

Fig. 3 The selection of candidate auxiliary objects

下载: 全尺寸图片幻灯片

图 4 目标位置区域的四种情况

Fig. 4 Four possible situations of the region

下载: 全尺寸图片幻灯片

图 5 颜色概率模型$P_{v}(I(x))$

Fig. 5 The color probability model

下载: 全尺寸图片幻灯片

图 6 融合示例

Fig. 6 An example of the integration model

下载: 全尺寸图片幻灯片

图 7 有效辅助目标数量变化曲线(Bolt2视频)

Fig. 7 The number curve of effective auxiliary objects

下载: 全尺寸图片幻灯片

图 8 Bolt2第10、180和290帧中有效辅助目标及其权值

Fig. 8 The effective auxiliary objects in frames 10, 180 and 290 of video Bolt2

下载: 全尺寸图片幻灯片

图 9 实时的中心位置误差曲线(BlurBody视频1 ~ 334帧)

Fig. 9 The real-time error curves of CLE in frames 1 to 334 of BlurBody video

下载: 全尺寸图片幻灯片

图 10 累计中心位置误差曲线(BlurBody视频1 ~ 334帧)

Fig. 10 The cumulative error curves of CLE in frames 1 to 334 of BlurBody video

下载: 全尺寸图片幻灯片

图 11 背景干扰时的跟踪结果

Fig. 11 The tracking results under background clutter

下载: 全尺寸图片幻灯片

图 12 目标形变时的跟踪结果

Fig. 12 The tracking results under deformation

下载: 全尺寸图片幻灯片

图 13 遮挡时的跟踪结果

Fig. 13 The tracking results under occlusion

下载: 全尺寸图片幻灯片

图 14 目标尺度突变时的跟踪结果

Fig. 14 The tracking results under scale variations

下载: 全尺寸图片幻灯片

表 1 辅助目标对跟踪速度(帧/秒)和中心位置误差(像素)的影响

Table 1 The effect of auxiliary objects number on speed (FPS) and precision (CLE)

辅助目标数量	KCFAOT		PFAOT
辅助目标数量	FPS	CLE	FPS	CLE
0	178.47	329.81	1.96	11.11
1	34.5	120.17	1.85	8.38
5	16.11	7.39	1.8	7.81
10	9.75	6.58	1.62	8.18
20	5.43	7.15	1.51	8.12
40	3.167	6.79	1.25	8.42

下载: 导出CSV

表 2 所提出方法与12种跟踪算法在50个视频下的中心位置误差阈值为20像素的平均成功率(%)和平均中心位置误差(像素)比较

Table 2 The performance of our method and 12 other trackers in 50 videos in terms of success rate (%) at 20 pixels threshold and mean center location error (pixels)

评价标准	KCFAOT	KCF	CT	Frag	DFT	IVT	CPF	ASLA	CXT	KMS	L1APG	LOT	TLD
平均成功率	76.29	68.67	35.16	38.53	34.69	48.89	53.52	61.06	56.68	42.16	51.65	44.12	53.97
平均位置误差	42.56	63.64	90.71	107.77	112.24	97.69	53.39	69.74	69.37	-	92.86	-	-
"-"表示有判断目标丢失

下载: 导出CSV

表 3 KCFAOT与12种跟踪方法的成功率(%)结果对比

Table 3 The overlap success rates (%) of KCFAOT and 12 other trackers

	KCFAOT	KCF	CT	Frag	DFT	IVT	CPF	ASLA	CXT	KMS	L1APG	LOT	TLD
Basketball	98	92	9	82	89	33	74	60	4	62	31	64	48
Biker	46	46	48	37	46	44	48	46	70	44	44	44	39
Bird1	35	7	29	7	28	3	39	22	6	31	4	21	0
BlurBody	72	58	2	27	12	2	14	1	46	38	87	10	1
BlurOwl	55	23	7	16	11	5	17	7	99	54	10	8	63
Bolt	100	99	1	14	4	2	96	2	3	34	2	90	1
Bolt2	94	2	100	45	2	3	89	2	5	61	2	59	1
BlurCar3	92	99	22	54	11	17	14	11	100	15	32	29	78
BlurCar4	91	99	3	47	43	3	12	30	100	54	99	13	11
Board	62	68	5	50	9	5	26	17	4	48	4	10	8
Boy	100	100	66	57	49	33	100	44	96	99	93	67	100
Car2	99	100	42	39	13	100	80	100	100	7	100	8	92
Car24	100	100	55	21	11	100	60	100	100	17	100	39	97
CarDark	100	100	2	55	54	71	23	100	70	63	100	63	65
CarScale	79	81	64	68	65	78	67	73	70	56	65	47	63
Couple	74	26	31	91	9	9	85	23	64	11	61	64	22
Coupon	100	100	17	21	100	100	21	100	100	37	15	28	15
Crossing	100	100	100	40	68	100	88	100	56	100	25	41	57
Crowds	100	100	1	3	100	12	3	100	100	2	100	2	59
Dancer2	96	100	93	84	100	100	77	100	100	67	100	61	100
David	100	100	82	10	31	100	23	100	100	49	81	37	66
David3	96	100	40	79	75	75	55	72	16	98	46	98	35
Deer	99	82	1	15	31	3	7	4	80	54	72	25	28
Diving	57	54	5	18	25	30	73	37	20	75	20	46	19
Dog	100	99	99	86	72	15	86	100	100	62	94	90	82
DragonBaby	53	34	22	48	12	33	49	29	72	49	22	70	18
FaceOcc1	75	78	9	98	62	65	31	53	42	54	65	26	17
Fish	91	100	87	56	86	100	12	100	100	31	5	21	40
Football	17	80	62	98	84	79	18	75	80	15	80	99	79
Freeman3	92	91	10	61	65	76	93	100	100	51	75	60	83
Girl	98	86	69	65	30	47	75	100	94	54	100	64	94
Girl2	84	7	7	16	7	8	97	38	11	80	7	79	25
Gym	84	79	27	94	23	61	95	78	63	97	2	96	85
Human3	3	1	2	2	1	1	2	1	1	2	1	2	1
Human4	52	53	20	8	20	19	30	20	11	22	19	25	15
Human5	45	27	24	4	9	28	12	100	38	24	96	61	99
Human6	34	29	26	33	28	38	47	52	21	26	43	37	48
Lemming	54	49	1	47	52	17	87	17	77	81	17	79	80
Matrix	33	17	14	7	6	2	18	18	12	24	12	1	1
MotorRolling	5	5	4	7	4	3	4	5	2	5	3	5	10
RedTeam	100	100	100	26	47	100	79	100	44	19	100	21	57
Skater	98	93	98	59	48	98	96	98	100	94	84	95	90
Skater2	91	69	1	57	6	7	50	25	17	52	6	74	32
Skating2	60	38	6	11	5	9	48	24	4	28	3	57	2
Skiing	64	7	11	4	7	11	6	14	9	11	9	2	6
Sylvester	81	84	48	74	44	68	83	80	100	66	52	82	91
Trellis	97	100	37	37	51	35	23	91	86	35	18	31	43
Twinnings	99	90	99	64	83	58	59	73	62	69	93	68	80
Walking2	47	43	40	35	40	100	36	65	44	48	98	39	38
Woman	98	94	20	19	95	20	19	94	14	13	20	16	40
效果最好的用粗体标注

下载: 导出CSV

表 4 KCFAOT与12种跟踪方法的平均位置误差(像素)结果对比

Table 4 The CLE (pixels) of KCFAOT and 12 other trackers

	KCFAOT	KCF	CT	Frag	DFT	IVT	CPF	ASLA	CXT	KMS	L1APG	LOT	TLD
Basketball	7.1	8.07	100.58	11.52	18.03	88.21	55.34	102.48	171.16	76.11	137.81	69.23	-
Biker	30.46	77.18	24.49	93.46	81.68	93.81	51.99	85.57	14.25	103.96	80.72	74.58	-
Bird1	118.05	151.82	123.15	325.31	131.37	175.61	122.24	124.07	212.19	-	334.25	170.4	-
BlurBody	16.03	64.03	87.12	37.79	259.16	167.37	45.91	145.89	22.74	27.33	12.05	94.57	-
BlurOwl	65.83	183.43	109.42	223.42	111.12	166.78	62.16	108.41	6.63	24.91	176.33	83.44	-
Bolt	7.63	6.74	379.41	258.83	367.28	377.62	12.43	390.85	376.59	109.46	408.53	13.68	-
Bolt2	7.15	329.81	9.68	43.21	276.7	82.69	10.8	284.67	237.22	39.93	304.19	78.84	-
BlurCar3	12.6	4.14	66.11	58.22	154.69	142.55	57.25	101.79	3.99	97.1	112.88	113.92	-
BlurCar4	12.16	10.72	158.49	56.8	18.53	131.6	45.31	48.79	10.12	21.88	7.04	183.2	-
Board	19.56	35.74	79.02	92.26	100.16	148.77	55.1	90.13	152.34	25.39	173.97	186.21	148.56
Boy	4.48	2.67	37.77	33.9	106.31	91.82	4.83	73.66	4.03	5.37	7.17	64.33	4.09
Car2	3.38	3.97	47.09	89.4	87.69	1.66	12.11	1.42	1.57	120.63	1.32	106.05	-
Car24	5.83	4.41	77.14	127.07	165.62	1.36	26.01	2.53	2.29	-	2.24	87.82	-
CarDark	5.86	5.76	121.4	37.72	58.85	19.29	42.13	0.98	21.63	28.78	1.16	28.49	-
CarScale	12.88	16.14	77.39	15.03	75.75	11.4	30.41	15.31	25.01	39.81	79.95	99.69	-
Couple	16.75	47.17	76.43	9.79	108.6	134.64	13.29	87.82	40.74	108.78	28.72	36.96	-
Coupon	1.44	1.34	20.66	71.57	2.52	11.8	72.06	1.88	4.69	24.11	33.53	26.15	-
Crossing	2.94	2.42	4.64	57.67	22.28	2.97	10.18	1.37	30.73	9.18	63.29	53.22	27.04
Crowds	4.91	3	413.86	371.33	3.33	236.36	369.35	4.26	4.44	364.21	4.61	394.52	-
Dancer2	7.61	6.41	8.33	11.24	5.51	8.38	13.22	6.31	7.64	17.35	8.02	15.76	6.89
David	4.52	8.06	15.1	93.02	42.88	4.78	26.05	5.33	7.96	19.44	13.98	23.99	-
David3	6.71	4.06	78.91	12.86	50.93	51.76	18.49	54.35	221.97	9.57	90.03	9.5	135.75
Deer	5.64	21.27	240.97	111.85	98.75	194.3	93.9	152.9	13.32	44.78	24.07	64.54	-
Diving	27.96	42.22	98.97	84.74	50.57	74.89	17.73	80.79	66.14	15.7	95.76	36.99	-
Dog	7.38	5.22	10.08	12.18	15.81	99.03	10.3	6.32	7.45	24.07	9.01	10.19	-
DragonBaby	74.78	50.52	53.69	46.39	75.58	92.86	35.56	49.88	20.31	-	101.62	27.88	-
FaceOcc1	15.82	14.83	36.1	11.17	23.59	18.74	27.8	37.74	22.62	18.97	17.82	34.5	33.48
Fish	9.41	4.08	9.39	22.73	8.84	5.09	39.16	3.81	5.93	29.3	29.64	34.16	-
Football	230.41	14.8	22.17	5.51	9.29	14.56	181.72	15.26	13.38	105.89	15.19	6.81	14.29
Freeman3	9.34	19.57	38.31	58.3	32.6	36.07	6.3	3.58	3.42	83.81	33.31	40.22	-
Girl	8.27	11.92	19.77	20.12	23.98	26.38	18.39	6.72	5.74	30.63	2.75	22.94	8.32
Girl2	35.29	264.55	103.47	257.63	236.34	209.21	8.31	86.98	135.45	14.68	220.53	34.13	-
Gym	15.16	16.45	25.65	10.04	104.47	27.52	10.52	14.83	19.18	9.56	79.91	9.69	14.23
Human3	241.45	260.8	212.6	224.23	227.09	262.45	157.83	252.25	173.66	254.84	284.02	207.78	-
Human4	130.56	131.46	295.71	310.21	245.91	295.95	202.23	308.54	327.55	-	319.63	212.91	-
Human5	112.98	175.52	238.86	286.2	259.06	180.92	218.33	6.8	188.88	-	7.47	89.44	-
Human6	83.42	107.67	53.24	84.19	131.11	115.03	23.11	79.22	87.48	205.11	63.6	43.05	-
Lemming	81.61	77.97	141.85	98.84	77.75	184	11.48	185.26	27.71	14	177.44	20.41	-
Matrix	144.89	76.42	98.9	163.58	105.78	97.26	136.88	55.14	162.7	50.51	62.01	-	-
MotorRolling	167.49	228.06	167.86	142.22	174.17	169.64	155.46	190.88	142.5	147.97	207.11	133.19	-
RedTeam	4.32	3.81	5.17	54.9	50.26	2.83	13.88	3.02	29.85	122.59	4.26	69.74	-
Skater	10.18	10.94	8.27	24.25	46.03	9.14	8.8	7.84	6.86	11.33	14.21	9.23	11.27
Skater2	10.48	18.07	63.11	20.02	125.04	63.86	23.09	37.08	56.95	24.57	110.4	15.71	-
Skating2	20.48	30.76	64.53	126.57	259.39	157.18	24.85	45.32	203.63	31.19	189.21	22.71	-
Skiing	31.22	260.45	265.01	271.52	276.2	276.76	256.2	252.66	145.09	265.3	265.73	248.51	-
Sylvester	13.65	13.3	37.08	14.4	44.88	36.75	14.16	19.71	5.79	18.24	25.97	11.68	9.88
Trellis	12.99	8.21	34.92	63.65	44.87	108.55	43.9	5.46	10.44	46.52	61.96	51.09	-
Twinnings	4.94	6.67	10.35	16.03	12.6	16.73	21.09	12.43	18.09	19.44	10.79	28.74	-
Walking2	26.27	29.57	62.06	57.31	29.09	4.19	55.02	29.91	34	44.98	4.7	64.56	-
Woman	6.22	10.09	120.16	106.69	8.5	188.89	120.59	10.22	120.05	130.24	128.51	130.56	-
效果最好的用粗体标注

下载: 导出CSV

表 5 测试视频属性统计

Table 5 The counts for the attitudes of test videos

视频属性	IV	SV	OCC	DEF	MB	FM	IPR	OPR	OV	BC	LR
视频数量	15	34	22	24	14	21	21	29	6	17	5

下载: 导出CSV

表 6 不同方法不同视频属性下的平均成功率(%)

Table 6 The average success rates (%) of different trackers in different videos

平均成功率	KCFAOT	KCF	CT	Frag	DFT	IVT	CPF	ASLA	CXT	KMS	L1APG	LOT	TLD
总计	76.29	68.67	35.16	38.53	34.69	48.89	53.52	61.06	56.68	42.16	51.65	44.12	53.97
IV	86.27	85.64	36.43	36.98	40.23	62.92	54.5	76.9	76.59	36.54	59.11	43.13	71.28
SV	73.94	64.03	36.38	31.38	27.18	48.47	55.68	60.9	54.58	40.05	52.75	41.64	56.48
OCC	65.84	56.02	30.46	33.22	37.51	42.21	51.68	54.18	35.99	40.28	44	44.97	41.74
DEF	65.83	46.98	22.04	29.19	28.03	24.8	49.21	46.09	26.11	42.76	28.87	47.63	34.96
MB	82.02	62.57	24.14	30.78	24.65	29.73	51.21	45.96	57.26	48.69	44.39	34.39	46.75
FM	69.1	61.26	20.95	37.89	30.96	26.79	50.49	38.34	55.32	44.66	38.94	36.61	49
IPR	83.75	80.97	48.82	46.96	36.87	56.44	64.21	68.05	68.99	50.34	56.92	51.96	62.85
OPR	73.19	64.25	36.65	43.41	38.45	42.12	61.68	57.6	47.64	49.64	41.33	52.14	52.08
OV	48.38	42.53	13.61	39.5	33.48	19.29	57.5	27.06	41.02	53.6	20.09	47.09	45.4
BC	75.23	73.75	32.04	32.97	29.24	56.03	42.38	66.21	59.2	26.01	55.89	32.2	53.7
LR	82.18	81.07	78.5	26.51	42.76	89.02	63.69	82.97	44.08	25.73	90.34	24.76	49.27

下载: 导出CSV

表 7 不同方法不同视频属性下的均位置误差(像素)

Table 7 The CLE (pixels) of different trackers in different videos

平均位置误差	KCFAOT	KCF	CT	Frag	DFT	IVT	CPF	ASLA	CXT	KMS	L1APG	LOT	TLD
总计	42.56	63.64	90.71	107.77	112.24	97.69	53.39	69.74	69.37	-	92.86	-	-
IV	27.28	28.34	101.56	106.46	95.07	79.25	56.26	56.55	48.16	-	70.23	-	-
SV	47.56	74.49	89.47	119.71	127.29	102.67	50.42	69.81	68.08	-	94.76	-	-
OCC	67.08	92.03	108.12	127.01	122.41	129.72	65.4	97.11	107.06	-	126.09	-	-
DEF	68.53	119.62	134.71	158.55	161.94	163.45	83.07	114.44	139.42	-	162.43	94.91	-
MB	25.08	91.08	84.77	129.95	124.54	131.5	40.86	68.78	62.31	-	108.69	91.88	-
FM	41.7	58.17	89.35	99.84	108.15	126.23	43.05	84.66	63.31	-	112.59	-	-
IPR	26.33	30.69	55.28	63.16	83.94	66.27	32.75	47.52	34.93	-	60.3	-	-
OPR	49.95	70.87	73.77	88.79	99.27	105.91	42.55	71.82	72.25	-	105.91	-	-
OV	71.56	83.98	99.35	118.76	100.71	153.69	38.23	126.56	87	-	162.85	78.7	-
BC	58	64.08	105.81	117.38	123.86	86.74	73.6	72.61	71.5	-	90.97	-	-
LR	19.19	25.75	31.87	63.86	56.57	22.35	33.52	26.85	35.99	107.27	20.63	72.65	-

下载: 导出CSV

参考文献(50)

[1]	Yilmaz A, Javed O, Shah M. Object tracking:a survey. ACM Computing Surveys, 2006, 38(4):Article No.13 doi: 10.1145/1177352
[2]	Li X, Hu W M, Shen C H, Zhang Z F, Dick A, van den Hengel A. A survey of appearance models in visual object tracking. ACM Transactions on Intelligent Systems and Technology, 2013, 4(4):Article No.58
[3]	Liu Q, Zhao X G, Hou Z G. Survey of single-target visual tracking methods based on online learning. Computer Vision, 2014, 8(5):419-428 doi: 10.1049/iet-cvi.2013.0134
[4]	Smeulders A, Chu D M, Cucchiara R. Visual tracking:an experimental survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(7):1442-1468 doi: 10.1109/TPAMI.2013.230
[5]	尹宏鹏, 陈波, 柴毅, 刘兆栋.基于视觉的目标检测与跟踪综述.自动化学报, 2016, 42(10):1466-1489 http://www.aas.net.cn/CN/abstract/abstract18935.shtml Yin Hong-Peng, Chen Bo, Chai Yi, Liu Zhao-Dong. Vision-based object detection and tracking:a review. Acta Electronica Sinica, 2016, 42(10):1466-1489 http://www.aas.net.cn/CN/abstract/abstract18935.shtml
[6]	Comaniciu D, Ramesh V, Meer P. Real-time tracking of non-rigid objects using mean shift. In:Proceedings of the 2000 IEEE Conference on Computer Vision and Pattern Recognition. Hilton Head, SC, USA:IEEE, 2000. 2142-2151
[7]	Vojir T, Noskova J, Matas J. Robust scale-adaptive mean-shift for tracking. In:Proceedings of the 2013 Scandinavian Conference on Image Analysis. Berlin, Germany:Heidelberg Springer, 2013. 652-663
[8]	Zhou S K, Chellappa R, Moghaddam B. Visual tracking and recognition using appearance-adaptive models in particle filters. IEEE Transactions on Image Processing, 2004, 13(11):1491-1506 doi: 10.1109/TIP.2004.836152
[9]	Oron S, Bar-Hillel A, Levi D, Avidan S. Locally orderless tracking. International Journal of Computer Vision, 2015, 111(2):213-228 doi: 10.1007/s11263-014-0740-6
[10]	Delabarre B, Marchand E. Dense non-rigid visual tracking with a robust similarity function. In:Proceedings of the 2014 IEEE International Conference on Image Processing (ICIP). Paris, France:IEEE, 2014. 2-7
[11]	Lin L, Lu Y Y, Li C L, Cheng H, Zuo W M. Detection-free multi-object tracking by reconfigurable inference with bundle representations. IEEE Transactions on Cybernetics, 2016, 46(11):2447-2458 doi: 10.1109/TCYB.2015.2478515
[12]	Liu X B, Lin L, Yan S C, Jin H, Jiang W B. Adaptive object tracking by learning hybrid template online. IEEE Transactions on Circuits and Systems for Video Technology, 2011, 21(11):1588-1599 doi: 10.1109/TCSVT.2011.2129410
[13]	Ross D A, Lim J, Lin R S, Yang M H. Incremental learning for robust visual tracking. International Journal of Computer Vision, 2008, 77(1-3):125-141 doi: 10.1007/s11263-007-0075-7
[14]	Lu D, Li L S, Yan Q S. A survey:target tracking algorithm based on sparse representation. In:Proceedings of the 7th International Symposium on Computational Intelligence and Design. Hangzhou, China:IEEE, 2015. 195-199
[15]	Mei X, Ling H B. Robust visual tracking using ι minimization. In:Proceedings of the 12th IEEE International Conference on Computer Vision. Kyoto, Japan:IEEE, 2009. 1436-1443
[16]	Mei X, Ling H B, Wu Y, Blasch E, Bai L. Minimum error bounded efficient ι tracker with occlusion detection. In:Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition. Colorado Springs, CO, USA:IEEE, 2011. 1257-1264
[17]	Wang J Y, Chen X L, Gao W. Online selecting discriminative tracking features using particle filter. In:Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. San Diego, CA, USA:IEEE, 2005. 1037-1042
[18]	Babenko B, Yang M H, Belongie S. Visual tracking with online multiple instance learning. In:Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition. Miami, FL, USA:IEEE, 2009. 983-990
[19]	Kalal Z, Mikolajczyk K, Matas J. Tracking-learning-detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(7):1409-1422 doi: 10.1109/TPAMI.2011.239
[20]	王丽佳, 贾松敏, 李秀智, 王爽.基于改进在线多示例学习算法的机器人目标跟踪.自动化学报, 2014, 40(12):2916-2925 http://www.aas.net.cn/CN/abstract/abstract18569.shtml Wang Li-Jia, Jia Song-Min, Li Xiu-Zhi, Wang Shuang. Person following for mobile robot using improved multiple instance learning. Acta Automatica Sinica, 2014, 40(12):2916-2925 http://www.aas.net.cn/CN/abstract/abstract18569.shtml
[21]	Grabner H, Bischof H. On-line boosting and vision. In:Proceedings of the 2006 IEEE Conference on Computer Vision and Pattern Recognition. New York, NY, USA:IEEE, 2006. 260-267
[22]	Hare S, Golodetz S, Saffari A, Vineet V, Cheng M M, Hicks S L, Torr P H. Struck:structured output tracking with kernels. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2016, 38(10):2096-2109 doi: 10.1109/TPAMI.2015.2509974
[23]	Matas J, Mikolajczyk K, Kalal Z. P-N learning:bootstrapping binary classifiers by structural constraints. In:Proceedings of the 2010 IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, CA, USA:IEEE, 2010. 49-56
[24]	Wang X Y, Hua G, Han T X. Discriminative tracking by metric learning. In:Proceedings of the 11th European Conference on Computer Vision. Heraklion, Crete, Greece:Springer, 2010. 200-214
[25]	李庆武, 朱国庆, 周妍, 霍冠英.基于特征在线选择的目标压缩跟踪算法.自动化学报, 2015, 41(11):1961-1970 http://www.aas.net.cn/CN/abstract/abstract18770.shtml Li Qing-Wu, Zhu Guo-Qing, Zhou Yan, Huo Guan-Ying. Object compressive tracking via online feature selection. Acta Automatic Sinica, 2015, 41(11):1961-1970 http://www.aas.net.cn/CN/abstract/abstract18770.shtml
[26]	Wu Y, Lim J, Yang M H. Object tracking benchmark. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1834-1848 doi: 10.1109/TPAMI.2014.2388226
[27]	Yang M, Wu Y, Hua G. Context-aware visual tracking. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(7):1195-1209 doi: 10.1109/TPAMI.2008.146
[28]	Dinh T B, Vo N, Medioni G. Context tracker:exploring supporters and distracters in unconstrained environments. In:Proceedings of the 2011 IEEE Conference on Computer Vision and Pattern Recognition. Colorado Springs, CO, USA:IEEE, 2011. 1177-1184
[29]	Wen L Y, Cai Z W, Lei Z, Yi D, Li S Z. Online Spatio-temporal structural context learning for visual tracking. In:Proceedings of the 12th European Conference on Computer Vision. Florence, Italy:Springer, 2012. 716-729
[30]	Zhu G B, Wang J Q, Zhao C Y, Lu H Q. Weighted part context learning for visual tracking. IEEE Transactions on Image Processing, 2015, 24(12):5140-5151 doi: 10.1109/TIP.2015.2479460
[31]	Quan W, Jiang Y Q, Zhang J J, Chen J X. Robust object tracking with active context learning. The Visual Computer, 2015, 31(10):1307-1318 doi: 10.1007/s00371-014-1012-8
[32]	Li Y, Zhu J K, Hoi S C H. Reliable patch trackers:robust visual tracking by exploiting reliable patches. In:Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA:IEEE, 2015. 353-361
[33]	Yuan X H, Kong L B, Feng D C, Wei Z C. Automatic feature point detection and tracking of human actions in time-of-flight videos. IEEE/CAA Journal of Automatica Sinica, 2017, 4(4):677-685 doi: 10.1109/JAS.2017.7510625
[34]	Bolme D S, Beveridge J R, Draper B A, Lui Y M. Visual object tracking using adaptive correlation filters. In:Proceedings of the 2010 IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, CA, USA:IEEE, 2010. 2544-2550
[35]	Henriques J F, Caseiro R, Martins P, Batista J. Exploiting the circulant structure of tracking-by-detection with kernels. In:Proceedings of the 12th European Conference on Computer Vision. Florence, Italy:Springer, 2012. 702-715
[36]	Henriques J F, Rui C, Martins P, Batista J. High-speed tracking with kernelized correlation filters. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 37(3):583-596 http://www.ncbi.nlm.nih.gov/pubmed/26353263
[37]	Liu T, Wang G, Yang Q X. Real-time part-based visual tracking via adaptive correlation filters. In:Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA:IEEE, 2015. 4902-4912
[38]	Ma C, Yang X K, Zhang C Y, Yang M H. Long-term correlation tracking. In:Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA:IEEE, 2015. 5388-5396
[39]	Zhang K H, Zhang L, Liu Q S, Zhang D, Yang M H. Fast visual tracking via dense spatio-temporal context learning. In:Proceedings of the 13th European Conference on Computer Vision. Zurich, Switzerland:Springer, 2014. 127-141
[40]	Stuart R, Peter N[著], 姜哲, 金奕江, 张敏[译]. 人工智能: 一种现代方法(第2版). 北京: 人民邮电出版社, 2010. 368-370 Stuart R, Peter N[Author], Jiang Zhe, Jin Yi-Jiang, Zhang Min[Translator]. Artificial Intelligence:A Modern Approach (2nd Edition). Beijing:Posts and Telecommunications Press, 2010. 368-370
[41]	Jiang H L, Li J H, Wang D, Lu H C. Multi-feature tracking via adaptive weights. Neurocomputing, 2016, 207:189-201 doi: 10.1016/j.neucom.2016.03.074
[42]	Chen H K, Zhao X G, Tan M. Target salient confidence for visual tracking. In:Proceedings of the 7th International Congress on Image and Signal Processing. Dalian, China:IEEE, 2015. 436-441
[43]	Possegger H, Mauthner T, Bischof H. In defense of color-based model-free tracking. In:Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA:IEEE, 2015. 2113-2120
[44]	Zhang K H, Zhang L, Yang M H. Real-time compressive tracking. In:Proceedings of the 12th European Conference on Computer Vision. Florence, Italy:Springer, 2012. 864-877
[45]	Adam A, Rivlin E, Shimshoni I. Robust fragments-based tracking using the integral histogram. In:Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. New York, NY, USA:IEEE, 2006. 798-805
[46]	Sevilla-Lara L, Learned-Miller E. Distribution fields for tracking. In:Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence, RI, USA:IEEE, 2012. 1910-1917
[47]	Pérez P, Hue C, Vermaak J, Gangnet M. Color-based probabilistic tracking. In:Proceedings of the 7th European Conference on Computer Vision. Copenhagen, Denmark:Springer, 2002. 661-675
[48]	Jia X, Lu H C, Yang M H. Visual tracking via adaptive structural local sparse appearance model. In:Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence, RI, USA:IEEE, 2012. 1822-1829
[49]	Comaniciu D, Ramesh V, Meer P. Kernel-based object tracking. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2003, 25(5):564-575 doi: 10.1109/TPAMI.2003.1195991
[50]	Bao C L, Wu Y, Ling H B, Ji H. Real time robust L1 tracker using accelerated proximal gradient approach. In:Proceedings of the 2012 IEEE Conference on Computer Vision and Pattern Recognition. Providence, RI, USA:IEEE, 2012. 1830-1837