Feature-Weight and Scale Adaptive Algorithm for Kernel Correlation Tracking

Hongji Zhu; Fengqin Yu

doi:10.3788/LOP56.101501

Journals >Laser & Optoelectronics Progress >Volume 56 >Issue 10 >Page 101501 > Article

Laser & Optoelectronics Progress
Vol. 56, Issue 10, 101501 (2019)

Feature-Weight and Scale Adaptive Algorithm for Kernel Correlation Tracking

Hongji Zhu^* and Fengqin Yu

Author Affiliations

School of Internet of Things, Jiangnan University, Wuxi, Jiangsu 214122, China

show less

DOI: 10.3788/LOP56.101501 Cite this Article Set citation alerts

Hongji Zhu, Fengqin Yu. Feature-Weight and Scale Adaptive Algorithm for Kernel Correlation Tracking[J]. Laser & Optoelectronics Progress, 2019, 56(10): 101501 Copy Citation Text

show less

Tracking results obtained using different algorithms in 7 scenarios with different attributes. (a) Doll; (b) Skating 1; (c) Singer 2; (d) Jogging-2; (e) David 3; (f) Basketball; (g) Bolt

Fig. 1. Tracking results obtained using different algorithms in 7 scenarios with different attributes. (a) Doll; (b) Skating 1; (c) Singer 2; (d) Jogging-2; (e) David 3; (f) Basketball; (g) Bolt

Download full size

Fig. 2. Results of 51 video sequences. (a) Precision plot; (b) success rate plot

Download full size

Fig. 3. Test results of 28 video sequences with scale variation. (a) Precision plot; (b) success rate plot

Download full size

Algorithm	IV (25)	SV (28)	OCC (29)	DEF (19)	FM (17)	MB (12)	OV (6)	BC (21)	LR (4)	IPR (31)	OPR (39)
Proposed	0.748	0.733	0.834	0.814	0.561	0.548	0.617	0.733	0.584	0.764	0.796
SAMF	0.682	0.703	0.820	0.810	0.608	0.564	0.636	0.676	0.385	0.714	0.767
KCF	0.728	0.679	0.749	0.740	0.602	0.650	0.650	0.753	0.381	0.725	0.729
Struck	0.558	0.639	0.564	0.521	0.604	0.551	0.539	0.585	0.545	0.617	0.597
CN	0.603	0.618	0.642	0.637	0.530	0.604	0.499	0.612	0.490	0.677	0.661
TLD	0.537	0.606	0.563	0.512	0.551	0.518	0.576	0.428	0.349	0.584	0.596
CSK	0.481	0.503	0.500	0.476	0.381	0.342	0.379	0.585	0.411	0.547	0.540

Table 1. Tracking precision of 7 algorithms in 11 scenarios with different attributes

Algorithm	IV (25)	SV (28)	OCC (29)	DEF (19)	FM (17)	MB (12)	OV (6)	BC (21)	LR (4)	IPR (31)	OPR (39)
Proposed	0.678	0.613	0.750	0.785	0.558	0.553	0.633	0.681	0.524	0.690	0.702
SAMF	0.641	0.616	0.765	0.804	0.593	0.561	0.646	0.655	0.370	0.653	0.699
KCF	0.581	0.479	0.618	0.671	0.557	0.595	0.650	0.672	0.357	0.615	0.608
Struck	0.491	0.471	0.493	0.473	0.567	0.518	0.550	0.545	0.410	0.528	0.506
CN	0.484	0.440	0.509	0.553	0.501	0.557	0.496	0.536	0.447	0.555	0.524
TLD	0.460	0.494	0.468	0.456	0.473	0.482	0.516	0.491	0.327	0.476	0.497
CSK	0.388	0.352	0.404	0.370	0.380	0.336	0.410	0.388	0.397	0.457	0.439

Table 2. Tracking success rates of 7 algorithms in 11 different attribute scenarios

Hongji Zhu, Fengqin Yu. Feature-Weight and Scale Adaptive Algorithm for Kernel Correlation Tracking[J]. Laser & Optoelectronics Progress, 2019, 56(10): 101501

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information

微信扫一扫：分享

微信扫一扫：分享