Scribd
Upload
8 views10 pages

ComputerVision PPT Raunak

The document discusses thresholding and binarization techniques used to convert grayscale images into binary images, simplifying image analysis. It outlines different methods such as global, adaptive, and Otsu's thresholding, along with their applications and suitability for various lighting conditions. The importance of selecting the appropriate method based on image characteristics is emphasized for achieving accurate results in image processing tasks.

Uploaded by

EnRyuu Castadel
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
8 views10 pages

ComputerVision PPT Raunak

The document discusses thresholding and binarization techniques used to convert grayscale images into binary images, simplifying image analysis. It outlines different methods such as global, adaptive, and Otsu's thresholding, along with their applications and suitability for various lighting conditions. The importance of selecting the appropriate method based on image characteristics is emphasized for achieving accurate results in image processing tasks.

Uploaded by

EnRyuu Castadel
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Name: Raunak Saigal

University Roll : 10930622059


Class Roll: 59
Subject Name: Computer Vision
Subject Code: PECAIML701B
4th Year, 7th Semester
Stream: AIML
Definition:
Thresholding is a technique used to convert grayscale images
into binary (black and white) images by selecting a threshold
value

Purpose:
Simplifies image analysis by segmenting objects from the
background.

Applications:
Object detection, OCR, Medical imaging, License plate
recognition
Binarization:
The process of reducing a grayscale image to two levels (0 and 255), based on
a threshold.

Why Binarize?
Makes further processing like edge detection and object tracking simpler and
more efficient.
Concept: A single threshold value
is applied to the entire image.

Formula: If pixel value > T → 255


(white), else→ 0 (black).

Use Case: Images with uniform


lighting
Concept: Different threshold
values are computed for different
regions of the image based on
local properties.

Types: Mean adaptive, Gaussian


adaptive

Use Case: Images with varying


illumination.
Concept: Automatic thresholding
technique that minimizes intra-
class intensity variance.

How it works: Calculates the


optimal threshold by maximizing
between-class variance.

Use Case: When histogram is


bimodal (two peaks).
Method Threshold Type Suitable For Complexity

Global Fixed Uniform lighting Low

Adaptive Local Uneven illumination Medium

Otsu’s Automatic Bimodal histogram Medium-High


Global: Best for well-lit images with
good contrast.
Tip: Preprocessing (like
Adaptive: Ideal for documents or filtering) may enhance results
scenes with shadows.

Otsu’s: Great for images with clear


foreground/background separation.
Global Thresholding
Fails under uneven lighting
Adaptive Thresholding
May introduce noise

Otsu’s Method
Assumes bimodal histogram, which
may not always be the case
Summary Importance Closing Note
Thresholding and Choosing the right Understanding the
binarization are key method affects image characteristics
for image accuracy in is essential before
simplification. downstream tasks. choosing a technique.

You might also like