Improved kernel correlation filter tracking with Gaussian scale space

2016 ◽  
Author(s):  
Shukun Tan ◽  
Yunpeng Liu ◽  
Yicui Li
Keyword(s):  
Scale Space ◽  
Correlation Filter ◽  
Kernel Correlation ◽  
Gaussian Scale Space

scholarly journals FEATURES OF DERIVATIVE CONTINUITY IN SHAPE

1999 ◽  
Vol 13 (08) ◽  
pp. 1181-1199
Author(s):  
PHILIP F. HENSHAW
Keyword(s):  
Time Series ◽  
Image Recognition ◽  
Scale Space ◽  
Complex Processes ◽  
Behavioral Dynamics ◽  
Higher Derivative ◽  
Smoothing Kernel ◽  
Key Features ◽  
Invariant Relationship ◽  
Gaussian Scale Space

Derivative continuity is a distributed invariant relationship between parts of flowing shapes. The original techniques presented here were developed for making the behavioral dynamics of complex processes more recognizable, but are equally applicable to assisting in the recognition of shapes in images. Regularizing a sequence using a constraint of derivative continuity is equivalent to using a bimodal smoothing kernel, producing a distinct bias for reducing variation on higher derivative levels, sharply defining shape with minimal suppression of shape. To help determine where reconstructing shapes in this way is valid, a test was developed to help distinguish combinations of noise and smooth flows from random walks. This helps distinguish between illusory and genuine, data shapes but also exposes a flair in using this and other measures of scaling behavior for diagnostic purposes. Gaussian scale space techniques in use for some time in image recognition, for identifying reliable landmarks in the shapes of outlines, are demonstrated for use in identifying key features of shape in time series.

scholarly journals Qualitative Descriptions of Image in the Gaussian Scale-Space

2005 ◽  
Vol 11 (2) ◽  
pp. 157-166
Author(s):  
Tomoya SAKAI ◽  
Atsushi IMIYA
Keyword(s):  
Scale Space ◽  
Gaussian Scale Space

Fast motion tracking based on moth-flame optimization and kernel correlation filter

2020 ◽  
Vol 39 (3) ◽  
pp. 3825-3837
Author(s):  
Yibin Chen ◽  
Guohao Nie ◽  
Huanlong Zhang ◽  
Yuxing Feng ◽  
Guanglu Yang
Keyword(s):  
Motion Tracking ◽  
State Of The Art ◽  
Candidate Region ◽  
Correlation Filter ◽  
Data Sets ◽  
Search Window ◽  
Tracking Method ◽  
Fast Motion ◽  
Kernel Correlation

Kernel Correlation Filter (KCF) tracker has shown great potential on precision, robustness and efficiency. However, the candidate region used to train the correlation filter is fixed, so tracking is difficult when the target escapes from the search window due to fast motion. In this paper, an improved KCF is put forward for long-term tracking. At first, the moth-flame optimization (MFO) algorithm is introduced into tracking to search for lost target. Then, the candidate sample strategy of KCF tracking method is adjusted by MFO algorithm to make it has the capability of fast motion tracking. Finally, we use the conservative learning correlation filter to judge the moving state of the target, and combine the improved KCF tracker to form a unified tracking framework. The proposed algorithm is tested on a self-made dataset benchmark. Moreover, our method obtains scores for both the distance precision plot (0.891 and 0.842) and overlap success plots (0.631 and 0.601) on the OTB-2013 and OTB-2015 data sets, respectively. The results demonstrate the feasibility and effectiveness compared with the state-of-the-art methods, especially in dealing with fast or uncertain motion.

Sign in / Sign up

Export Citation Format

Share Document