Introduction:
The microscope in its various forms remains the most basic investigative tool of the Biologist. Light and electron images of biological specimens have been recorded on photographic emulsions which remain unsurpassed in their ability to provide the highest resolution. However, the capability of digitizing images and processing them with computer algorithms permits the enhancement of selected image features and also rapid and accurate mathematical analysis of cell and tissue structure. Enhancement of image information by digital processing has wide applications in biological microscopy and also in macroscopic realms such as photography, tomography, satellite imaging and astronomy. It is incumbent on serious investigators in these fields to develop an understanding of the fundamentals of computer image processing rather than simple cursory appreciation of a processed image looking better than the original.
Goals of the Course:
Graduate students in Biology should have a basic understanding of the principle methods and approaches of computer image processing and analysis as well as hands-on experience with using the computer to process and analyze microscopic images. This experience will enable them to employ this technology in their own experimental designs as well as to appreciate the methods of image processing appearing with ever-increasing frequency in the scientific literature.
Three Power MacIntosh 8600/200 workstations equipped with Scion Image Grabber boards and two separate image processing software packages (NIH Image 6.0 and Adobe Photoshop 4.0) will be utilized for this course. In addition, plug-in routines compatable with both Image and Photoshop along with image archives and tutorials for their use will be available for individual instruction in the application of computer processing and analysis of images.
Course Mechanics
Lectures will be held on Mond. afternoons and during the first half of the semester will cover the fundamental principles of computer image processing and analysis, human vision and biological microscopy. Weds. and Fri. labs will be spent practicing image capture methods, use of the Power MacInstosh 8600 computer, and familiarization with the NIH Image 6.0 and Adobe Photoshop 4.0 software and tutorials in the Image Processing Toolkit 2.1.
As with all computer applications, a distinct goal or project for the
use of computer software routines facilitates understanding and retention
of its principles to a greater degree than simply reading and trying various
software applications as set forth in the imaging routines. Consequently,
students will then be asked to choose a problem in microscopic analysis
and design methods to solve the problem using the capabilities of the hardware
and software available. The methods employed, the analytical protocols
used to solve the problem and an appraisal of the data generated will constitute
a final report to be presented formally to the class during the last half
of the course.
Lecture Schedule (tentative)
Sept. 3- Course Intro., PowerMac 8600, NIH Image 6.0
Adobe Photoshop 4.0, Image Processing Toolkit 2.1
Sept. 8- Image processing systems, Image aquisition, Digital
Images (bits & pixels), Image file formats
Sept. 15-Intensity transformations functions, Point operations
spatial convolution
Sept. 22-Segmentation and binary images
Sept. 29-Image analysis (global properties)
Oct. 6- Image Analysis (feature properties)
Oct. 13- Human Vision (Dr. Granda)
Oct. 20- fall break
Oct. 27- Light and Electron Microscopy
Nov. 3- Fourier analysis (Dr. Sharnoff)
Nov. 10- Deconvolution routines (Dr. Czymmek)
Nov. 17-Dec. 8 (student presentations)
I. Human Vision
A. The eye
1. Refraction and Accomodation
2. The retina
a. rods and cones
b. resolution and visual
acuity
c. photopic and scotopic
vision
B. Imaging
1. Object recognition
2. Illusions
3. Contrast Discrimination
4. Color Vision
5. Stereoscopy
II. Light Microscopy
A. Magnification
B. Resolution
C. Contrast
1. Amplitude and Phase Objects
D. Bright Field Microscopy
E. Phase Contrast Microscopy
F. Differential Interference Microscopy (DIC)
G. Fluorescence Microscopy
III. Electron Microscopy
A. Transmission and Scanning Electron Microscopy
IV. Image Sensors and Digital Image Aquisition
A. Photographic Emulsions
B. Video Cameras
V. Image Processing Systems
A. Optical Digitizer
1. Analog to digital Conversion
B. Image Processing System
C. Display
D. Digital Imaging (NIH Image routines)
1. Grey-level Histogram
2. Grey-level clipping
3. Contrast Range
4. Image Averaging
VI. Intensity Transformation Functions
A. Point Operations
1. Histogram Equalization
2. Look-up Tables (LUTS)
a. Linear and Logarithmic
Functions
b. Color Luts
B. Spatial Convolution
1. Convolution Masks (Kernals)
2. Spatial Frequency
3. Smoothing Filters
4. Rank Filters (minimum, medium, maximum)
5. Sharpening Filters
VII. Segmentation
A. Edge Detectors (Sobel and Kirsh Operators)
B. Shadow Filters
C. Discrimination and Thresholding
1. Density Slice
2. Thresholding Processed Images
D. Manual Outlining
VIII. Binary Images
A. Binary Image Editing
1. Manual Editing
2. Boolean Logic Operators
3. Neighbor Operations
a. Erosion and Dilation
b. Opening and Closing
4. Skeletonization
B. Fractal Dimension
IX. Image Measurements
A. Spatial Calibration
B. Global Parameters
1. Area Fraction
2. Number of Features
C. Feature Measurements
D. Brightness Measurements
X. Stereology
A. Principle of Delisse
B. Quantitation of Structural Parameters
1. Global Measurements
a. Volume Density
b. Surface Density
c. Length Density
2. Feature-Specific Measurements
a. Spheres
b. cylinders and ellipsoids
3. Effect of Section Thickness
a. Holmes Effect
C. Sampling Schemes
1. Randomly-Dispersed Samples
2. Unbiased Sampling of Fields
3. Layered Structures
4. Fasiculated Structures
5. Sampling Procedures
D. Statistical Analysis
XI. Three-Dimensional Reconstruction
A. Stereomicroscopy
B. Serial Section Reconstruction
1. Section Thickness
2. Stains
3. Computerized Reconstruction of Serial Images
XII. Fourier Transforms (Dr. Sharnoff)