Image Thresholding

Image Thresholding simply means setting an intensity value for intensity value lower than a certain value(Threshold value) and another intensity value for intensity value higher than the threshold value.

Here, Threshold value acts as a point of intensity value which separates the intensity values of pixels into two parts.

Threshold value can be selected by different methods and Image Thresholding can also be done by different ways. In this context, we are only taking a grayscale image(or monochrome image).

Some basic Image Thresholding techniques are:

a) Global Thresholding:

We select the Threshold value manually which seems suitable for the given picture.

Here, the selected Threshold value is applied for all pixels or whole image.

It can be implemented as:

if pixel_intensity >= Threshold : pixel_intensity = max_intensity

else: pixel_intensity = min_intensity

left(grayscale image before thresholding) right(image after Global Thresholding)

b) Mean Global Thresholding:

It is similar to Global Thresholding except the selection of Threshold value.

Here, the Threshold value is selected by calculating the mean intensity value from the whole image.

c) Mean Adaptive Thresholding:

Adaptive Thresholding includes adaptive Threshold value. Here, the Threshold value is selected or calculated for smaller regions or windows in the image giving different Threshold values for different regions of the image.

In case of selecting or calculating the Threshold value within a certain region or window, the mean intensity value is calculated from the intensity values of that region or window.

Then, the pixels intensity values within that region or window are selected on the basis of calculated Threshold value.

left(grayscale image before thresholding) right(image after Mean Adaptive Thresholding)