You are on page 1of 2

ADAPTIVE THRESHOLDING

The performance of shot detection methods in video sequences can be improved by


the use of a threshold that adapts itself to the sequence statistics. In many of the methods, the
detection decision is based on a hard threshold of some dissimilarity measure, whose value is
determined by experimentation. The optimal value depends on the requirements of the
application and will be a trade-off between the number of false positives detected and the
number of undetected true positives. One can use an adaptive threshold setting.

5.1 Using Mean and Variance:


A method was proposed in which the means and standard deviations for the whole
sample are calculated. Mean (μ) and the standard deviation (σ ) of the feature vector
differences as follows:
N−1
1
µ= ∑ D(i ,i¿+1)¿ (6.1)
N −1 i=1

N−1
σ=
√ 1
∑ [ D (i , i+1 ) −μ ]
N −1 i =1
(6.2)

Where, N is the number of frames in the video sequence and D (i,i+1) is difference
value.

Declare the frame I as boundary detection if

D (i,i+1)> μ + kσ (6.3)

Where, k is a pre-specified constant.

Algorithm:

1. Means and Standard Deviations of difference values are calculated.Meanμ and the
standard deviation σ of the feature vector distances as follows:
N−1
1
µ= ∑ D(i ,i¿+1)¿
N −1 i=1
N−1
σ=
√ 1
∑ [ D (i , i+1 ) −μ ]
N −1 i =1
Where, N is the number of frames in the video sequence and D (i, i+1) is
difference value.

2. Declare the frame I as boundary detection if

D ( i, i+ 1 )> μ+ kσ

where k is pre-specified constant.

You might also like