Walsh–Hadamard-Kernel-Based Features in Particle Filter Framework for Underwater Object Tracking
Citations
35 citations
3 citations
3 citations
2 citations
2 citations
References
8,738 citations
"Walsh–Hadamard-Kernel-Based Feature..." refers background in this paper
...4) Range Strength: The range measure indicates the photometric similarity between a pixel and its neighborhood [36]....
[...]
3,828 citations
"Walsh–Hadamard-Kernel-Based Feature..." refers methods in this paper
...The results are reported and analyzed using three performance measures: CLE, ATE, and OIoU indices and found to be performing the best....
[...]
...In this article, the evaluation measures used for the quantitative analysis of the proposed tracking technique are overall intersection over union (OIoU) [2], centroid location error (CLE) [14], and average tracking error (ATE) [2]....
[...]
...The ATE values are reported in Tables II–IV....
[...]
1,812 citations
1,631 citations
"Walsh–Hadamard-Kernel-Based Feature..." refers background or methods in this paper
...frt+ 1 = θFt + (1 − θ)fr1 (21) where θ is the update regulating parameter, which is equal to the Sφrensen distance [30] between the histogram of the rectangular blob at time t = 0 and t....
[...]
...Hence, in the proposed scheme, reference feature strengths are updated regularly on the basis of Sφrensen distance [30] between the histogram of the tracked blob at any time instant t and (t = 0) of subsequent frames....
[...]
...To compute the likelihood, the reference feature strengths are updated based on the Sφrensen distance [30]....
[...]
1,604 citations
"Walsh–Hadamard-Kernel-Based Feature..." refers background in this paper
...The position noise and velocity noise variance for the particle filter are manually fixed between [3, 7], and [1, 4], respectively, for different videos depending on the speed of object motion, speed of changing the shape with respect to the field of view....
[...]