Question

Kindly explain the following in detail:

1. Why is interpolation needed in the rotation process?

2. How do the different interpolation methods "nearest neighbor," "bilinear interpolation," and "bicubic interpolation" differ?

Answer 1

ANSWERS :

ANS 1. Interpolation is basically a method of evaluating unknown values that usually comes between the known values.

Interpolation is needed in rotation process because :

* It simplify the various complicated functions by providing data points to evaluate.

* To deal with the discrete data that is very difficuilt to solve and derive the result from.

* To maintain that the 90 degree rotation is losless,any further rotation can distort and affect the picture quality and pixels.

* In image rotation there can be many discrete data ,so in order to find an estimate value we use interpolation in image rotation.

ANS 2. The different interpolation are following :

a) Nearest Neighbour

b) bilinear interpolation

c) bicubic interpolation

* Nearest Neighbour :

- It is the basic form of interpolation.

- Also known as proximal interpolation.

- In this interpolation we do not calculate the average value but instead uses the nearest pixel value for the output pixel value.

* Bilinear interpolation :

- It is an extention of the nearest neighbour interpolation.

- It is ussually on a 2D grid

- In this nearest neighbour interpolation is used first in one direction and then in all the remaining direction one after the other.

- It takes only 4 pixels in account.

* Bicubic interpolation :

- It is an appendex of the cubic interpolation.

- It is also on a 2D regular grid.

- The interpolated surface is comparitively smoother than both Bilinear and nearest point interpolarity.

- It takes into consideration 16 pixels image.