Name

im_diff - Calculate scaled density difference between two images.

Usage

imdiff, A, B = im_diff(image1,image2,[optional: mask, or radius of a circular mask])

or use a tuple to hold the returned values.

S =im_diff(image1,image2,[optional: mask or radius of a circular mask])

• S[0] = imdiff
• S[1] = A
• S[2] = B

Input

image1

The first input image for calculating density difference. It can be either a read-in-core EMData object or a string of file in disk.The image dimension can be 1,2, or 3.

image2

The second input image for calculating density difference. It can be either a read-in-core EMData object or a string of file name in disk. The image dimension can be 1, 2 or 3.

mask

Mask file, the output density difference image will be calculated under this mask. It can be string of file name in the desk, read-in-core EMData object, integer or float number denoting radius of the circular mask.

• Note: The mask should be of the same dimensions as the input image.
• The default mask is a circle, or sphere, with the radius of min(nx/2-1, ny/2-1, nz/2-1)

Output

imdiff

image of the density difference between image1 and image2.

A

The linear error scale.

B

The average of total error between image1 and image2.

Description

1. `A = (N*sumP_1*P_2 -sumP_1*sumP_2 )/(N*sum(P_1)^2 -( sumP_1 )^2`

   • `P_1` and `P_2` are pixel values in image1 and image2, respectively.

   • Summation is over all the pixels under the given mask. N is the total number of pixels under mask.

2. `B = (A*sumP_1 -sumP_2)/N`

3. `"""imdiff""" = A*P_1-B -P_2`

4. When two images are identical, A = 1, B = 0.
5. Swapping image1 and image2 will give different A, B, and imdiff.

Author / Maintainer

Bharath K Narayanan/Zhong Huang

Keywords

category 1

UTILITIES

category 2

SPATIAL

Files

• statistics.py

Maturity

stable

works for most people, has been tested; test cases/examples available.

Bugs

None. It is perfect.