Perfect 👍 — now I’ll give you a HIGHLY DETAILED, POINTWISE, EXAM-WRITING
VERSION (you can directly write in answer sheet for 10 marks).
1) IMAGE ENHANCEMENT & HISTOGRAM PROCESSING
🔹 Image Enhancement
Definition
Image enhancement is the process of improving the visual quality of an image.
It enhances important features and suppresses unwanted information.
Objectives
Improve contrast
Highlight edges and details
Remove noise
Make image suitable for analysis
Key Techniques (Pointwise)
1. Image Negative
Formula: s = L − 1 − r
Converts bright → dark and vice versa
Useful in medical imaging (X-rays)
2. Log Transformation
Formula: s = c log(1 + r)
Expands dark regions
Compresses bright regions
3. Power Law (Gamma Correction)
Formula: s = crγ
Controls brightness
If γ < 1 → brighter image
If γ > 1 → darker image
4. Contrast Stretching
Expands intensity range
Improves low contrast images
Features
Works on pixel intensity
Improves visual perception
Simple preprocessing step
Advantages
Easy implementation
Enhances visibility
Useful in preprocessing
Disadvantages
Cannot recover lost data
May enhance noise
Applications
Medical imaging
Satellite images
Digital photography
🔹 Histogram Processing
Definition
Histogram = graphical representation of pixel intensity distribution.
Types (Important)
1. Histogram Equalization
Redistributes intensity uniformly
Improves global contrast
2. Histogram Matching
Matches histogram of one image to another
Features
Automatic enhancement
Global technique
Advantages
Improves contrast
Easy to apply
Disadvantages
Over-enhancement
Noise amplification
Applications
Face recognition
Medical image enhancement
2) IMAGE ENHANCEMENT IN SPATIAL DOMAIN
🔹 Spatial Domain
Definition
Processing directly on image pixels.
🔹 Spatial Filtering
Definition
Modify pixel values using a mask/kernel.
Working Steps
1. Place mask over image
2. Multiply mask values with pixels
3. Sum the result
4. Replace center pixel
Types
✅ Smoothing Filters
Mean Filter
Replaces pixel with average
Removes noise
Gaussian Filter
Weighted average
Better smoothing
✅ Sharpening Filters
Laplacian Filter
Detects edges
Second derivative
Sobel Operator
Detects horizontal & vertical edges
Features
Local processing
Uses convolution
Advantages
Simple
Fast computation
Disadvantages
Blurring in smoothing
Noise amplification in sharpening
Applications
Edge detection
Noise removal
3) FOURIER TRANSFORM & DFT
🔹 Fourier Transform
Definition
Converts spatial domain → frequency domain.
M −1 N −1
F (u, v) = ∑x=0 ∑y=0 f (x, y)e−j2π(ux/M +vy/N )
Concept
Low frequency → smooth areas
High frequency → edges/details
Features
Global transformation
Frequency separation
Advantages
Effective for filtering
Removes periodic noise
Disadvantages
Complex
Hard to interpret
🔹 DFT (Discrete Fourier Transform)
Definition
Digital version of Fourier Transform.
Features
Works on discrete signals
Used in computers
Advantages
Accurate frequency analysis
FFT makes it fast
Disadvantages
High computation (without FFT)
Applications
Image compression
Signal processing
4) COLOUR MODELS
🔹 Definition
Mathematical model to represent colors.
🔹 Types (Important for Exam)
RGB
Red, Green, Blue
Used in displays
CMY / CMYK
Used in printers
HSV
Hue → color
Saturation → purity
Value → brightness
YUV / YIQ
Used in TV/video
Features
Device dependent/independent
Converts color formats
Advantages
Efficient representation
Useful in image processing
Disadvantages
Conversion complexity
RGB device dependent
Applications
Graphics
Image editing
5) IMAGE COMPRESSION
🔹 Definition
Reduces image size without losing important information.
🔹 Types
Lossless
No data loss
Example: PNG
Lossy
Some data loss
Example: JPEG
🔹 Techniques
RLE
Stores repeated values
Huffman Coding
Variable-length coding
DCT
Used in JPEG
Features
Reduces storage
Faster transmission
Advantages
Saves memory
Reduces bandwidth
Disadvantages
Quality loss (lossy)
Processing time
Applications
Web images
Multimedia
6) NOISE MODELS & NOISE FILTERS
🔹 Noise Models
Types:
Gaussian → random noise
Salt & Pepper → black & white dots
Poisson → photon noise
Speckle → multiplicative noise
Effects
Degrades image quality
Affects analysis
🔹 Noise Filters
Types:
Mean Filter
Smooths noise
Median Filter
Best for salt & pepper
Gaussian Filter
Reduces Gaussian noise
Advantages
Removes noise
Improves clarity
Disadvantages
Loss of detail
Blurring
7) IMAGE DEGRADATION & RESTORATION
🔹 Degradation
Causes:
Noise
Blur
Motion
Model:
g(x, y) = h(x, y) ∗ f (x, y) + n(x, y)
🔹 Restoration
Methods:
Inverse Filtering
Removes blur
Wiener Filtering
Reduces noise + blur
Features
Mathematical model-based
Improves image quality
Advantages
Recovers original image
Accurate
Disadvantages
Complex
Requires prior knowledge
8) SEGMENTATION & APPLICATIONS
🔹 Segmentation
Definition
Divides image into meaningful regions.
Methods
Thresholding
Based on intensity
Edge Detection
Detects boundaries
Region Growing
Groups similar pixels
Features
Simplifies image
Important for analysis
Advantages
Easy object detection
Useful in AI
Disadvantages
Sensitive to noise
Complex in real cases
Applications
Medical imaging (tumor detection)
Face recognition
Autonomous vehicles
🔥 ULTRA FINAL EXAM REVISION (WRITE THIS FIRST)
👉 Most Important Topics:
Histogram Equalization
Spatial Filtering (Mean vs Median)
Fourier Transform
Compression (Lossy vs Lossless)
Noise types + filters
Segmentation methods