Image Processing Weeks 14 & 15

Image Processing - Week 14&15

• The topics covered in week 14 and 15 include color image processing and other related subjects.
• Final project presentations are scheduled for week 16.
• The presentation will be worth 6% of the total grade.
• The template for the presentation is available in the appendices.
• The presentation will consist of 10 minutes of presentation and 5 minutes for questions.

Color Fundamentals

• The human eye perceives thousands of color shades and intensities compared to the fewer shades of gray.
• The visible spectrum of light ranges from Gamma rays to Radio waves.

Color Models

• Color models are mathematical systems to represent color. The CIE 1931 Standard (International Commission on Illumination) is a common color model.
• This model uses wavelengths in nanometers.
• Red = 700nm
• Green = 546.1nm
• Blue = 435.8nm

Color Models and Notes

• The primary colors of light (RGB) are additive.
• The subtractive color model (CMY) is used for printing inks.
• RGB color model is used in displays like TVs and computer monitors.
• CMYK is used in print media like printers.

Color Maturity

• Color maturity has three elements: hue, lightness, and saturation.
• Hue refers to color properties (e.g., red, blue, green).
• Lightness describes the brightness of a color.
• Saturation describes the vividness of a color; high values are for bright and strong colors.

HSI Model

• The HSI color model is based on human perception of color.
• Hue is a subjective measure of color. Humans perceive roughly 200 different colors.
• Saturation is the relative purity of a color. Adding white to a color reduces saturation.
• Intensity refers to the brightness or darkness of an object.

Converting Between RGB and HSI

• Converting from RGB to HSI allows for better image manipulation of color.
• The intensity (I) of a color in the RGB model is calculated as (R+G+B)/3.
• Saturation (S) is calculated using this formula: S = 1 - min(R,G,B)/I
• Chroma is defined as max(R,G,B)-min(R,G,B).
• Hue (H) is determined based on the maximum color channel.

Pseudo-color Image Processing

• Pseudo-color is also known as false color.
• It assigns colors to gray-level values based on criteria (e.g., intensity slicing).
• This technique is used to improve human visual interpretation of grayscale images.
• Intensity slicing and gray levels to color transformations are used.

Image Segmentation

• Image segmentation partitions a digital image into segments (sets of pixels) to simplify analysis.
• This technique makes images easier to interpret and analyze.
• Common applications include autonomous vehicles and agriculture.
• Edge-Based Segmentation, Threshold-Based Segmentation, Region-Based Segmentation, Cluster-Based Segmentation, and Watershed Segmentation are different types.
• Intensity changes are useful for detecting edges in images.

Edge Detection

• Edge detection is a method used in image segmentation, based on changes in local intensities within an image.
• Techniques employ first and second derivative results.
• Some of the issues involved include the presence of noise which can create erroneous results. Filtering is often done before edge detection to mitigate the noise.
• Three typical edge types are: Step Edge, Ramp Edge, Roof Edge

Point and Line Detection

• Point and line detection techniques are methods that use image derivatives to find specific points and lines in an image.

Colorization

• Colorization transforms grayscale images to color images using algorithms based on intensity and texture analysis.