Image Analysis on Fiji

Transcript

1. IMAGE ANALYSIS ON FIJI Hiro Kai Zoom out (Ctrl+- (Windows) or ⌘‐ (Mac)) if you see cropped slides. 0
2. CONTENTS Introduction The basics of Fiji scripting A few important things in Python Using more ImageJ functions Real-world examples Advanced topics
3. INTRODUCTION
4. ABOUT THIS TUTORIAL I'll focus on practical issues of image processing on Fiji. Choice of content is flexible. Let me know if you have anything you want to work on. All source code I wrote for the tutorial is public domain. No credit, no copyright, free for modification. Any question is welcome. for resources and data for the tutorial Links
5. BATCH ANALYSIS BY SCRIPTING: MOTIVATION ImageJ/Fiji has some basic batch analysis functionalities. Using Python on Fiji, you can make even more sophisticated batch analysis.
6. COMPARISON BETWEEN MATLAB AND IMAGEJ Software Advantage Disadvantage Matlab Powerful matrix operation. Fast execution when coded appropiately Expensive, awkward design of languages and libraries when you want to write a well organized program. ImageJ (Fiji) Completely free, many readymade libraries, many scripting languages (Python, Ruby, Java, etc.) available Not many learning resources. Not many practical examples of scripts.
7. CHECKLIST FOR PYTHON'S LANGUAGE FEATURES Variables Lists i f /e l s e statement f o r loop Function definition by d e f keyword Module import Classes and objects
8. WORKFLOW OF IMAGE ANALYSIS 1. Detect particles, cells, etc. (by plugins). 2. Extract and summarize quantities from the detected objects. Custom Python code External programs , etc.
10. HOW TO USE PYTHON ON FIJI 1. Install Fiji Fiji comes with a Python (Jython) scripting engine. Fiji is written in Java language, and Jython is a Python environment that works with Java 2. Plugins → Scripting → Script Editor 3. Language → Python
11. RESOURCES FOR LEARNING Image processing algorithms in general Fiji Scripting Simple examples in Script Editor: Templates → Python Python/Jython Image analysis for biology Principles of Digital Image Processing Jython Scripting A Fiji Scripting Tutorial More examples Jython library reference
12. THE BASICS OF FIJI SCRIPTING
13. THE FIRST EXAMPLE OF FIJI SCRIPTING THRESHOLDING i m p = I J . o p e n I m a g e ( " h t t p : / / r s b . i n f o . n i h . g o v / i j / i m a g e s / b l o b s . g i f " ) i p = i m p . g e t P r o c e s s o r ( ) i p . s e t T h r e s h o l d ( 1 4 7 , 1 4 7 , I m a g e P r o c e s s o r . N O _ L U T _ U P D A T E ) I J . r u n ( i m p , " C o n v e r t t o M a s k " , " " ) i m p . s h o w ( )
14. WHAT IT DOES Open an image to get an I m a g e P l u s object, and assign it to i m p variable. Get an I m a g e P r o c e s s o r from the I m a g e P l u s Set threshold to 147, and convert the image to a binary image with it. Show the image. i m p = I J . o p e n I m a g e ( " h t t p : / / r s b . i n f o . n i h . g o v / i j / i m a g e s / b l o b s . g i f " ) i p = i m p . g e t P r o c e s s o r ( ) i p . s e t T h r e s h o l d ( 1 4 7 , 1 4 7 , I m a g e P r o c e s s o r . N O _ L U T _ U P D A T E ) I J . r u n ( i m p , " C o n v e r t t o M a s k " , " " ) i m p . s h o w ( )
15. IMAGES IN IMAGEJ
16. IMAGEPLUS KEEPS THE INFO ("METADATA") OF THE IMAGE Code from http://www.ini.uzh.ch/~acardona/fiji- tutorial/#imageplus # G r a b t h e l a s t a c t i v a t e d i m a g e a s a n I m a g e P l u s o b j e c t i m p = I J . g e t I m a g e ( ) p r i n t " t i t l e : " , i m p . t i t l e p r i n t " w i d t h : " , i m p . w i d t h p r i n t " h e i g h t : " , i m p . h e i g h t p r i n t " n u m b e r o f p i x e l s : " , i m p . w i d t h * i m p . h e i g h t p r i n t " n u m b e r o f s l i c e s : " , i m p . g e t N S l i c e s ( ) p r i n t " n u m b e r o f c h a n n e l s : " , i m p . g e t N C h a n n e l s ( ) p r i n t " n u m b e r o f t i m e f r a m e s : " , i m p . g e t N F r a m e s ( ) t y p e s = { I m a g e P l u s . C O L O R _ R G B : " R G B " , I m a g e P l u s . G R A Y 8 : " 8 - b i t " , I m a g e P l u s . G R A Y 1 6 : " 1 6 - b i t " , I m a g e P l u s . G R A Y 3 2 : " 3 2 - b i t " , I m a g e P l u s . C O L O R _ 2 5 6 : " 8 - b i t c o l o r " } p r i n t " i m a g e t y p e : " , t y p e s [ i m p . t y p e ]
17. THREES COMMON WAYS OF GENERATING AN IMAGEPLUS OBJECT # C a s e 1 : G e t t i n g a c u r r e n t i m a g e . i m p = I J . g e t I m a g e ( ) i m p . s h o w ( ) # S h o w t h e i m a g e j u s t t o m a k e s u r e i t ' s w o r k i n g . # C a s e 2 : R e a d i n g f r o m a U R L . i m p = I m a g e P l u s ( ' h t t p s : / / w w w . g o o g l e . c o m / i m a g e s / s r p r / l o g o 1 1 w . p n g ' ) i m p . s h o w ( ) # S h o w t h e i m a g e j u s t t o m a k e s u r e i t ' s w o r k i n g . # C a s e 3 : I n i t i a l i z i n g w i t h a n I m a g e P r o c e s s o r i m p = I m a g e P l u s ( ' h t t p s : / / w w w . g o o g l e . c o m / i m a g e s / s r p r / l o g o 1 1 w . p n g ' ) i p = i m p . g e t P r o c e s s o r ( ) # G e t a n I m a g e P r o c e s s o r f r o m t h e I m a g e P l u s i p 2 = i p . r o t a t e R i g h t ( ) # T h i s m a k e s a n e w I m a g e P r o c e s s o r o b j e c t i m p 2 = I m a g e P l u s ( ' M y r o t a t e d i m a g e ' , i p 2 ) # W r a p t h e I m a g e P r o c e s s o r w i t h I m a g e P l u s i m p 2 . s h o w ( )
18. IMAGEPROCESSOR HAS PIXEL DATA i m p = I J . g e t I m a g e ( ) # G r a b a n I m a g e P r o c e s s o r i p = i m p . g e t P r o c e s s o r ( ) # O b t a i n p i x e l s a s a l i s t . p i x e l s = i p . g e t P i x e l s ( ) # P r i n t t h e t o t a l a n d m e a n p i x e l i n t e n s i t i e s . s = s u m ( p i x e l s ) p r i n t s , 1 . 0 * s / l e n ( p i x e l s ) # A c c e s s e a c h p i x e l b y a c o o r d i n a t e . p r i n t i p . g e t P i x e l ( 1 0 0 , 1 2 0 ) # L i n e s c a n f o r p o i n t s ( x , 1 0 4 ) , x = 0 , 1 , 2 , . . . 9 9 x s = r a n g e ( 0 , 1 0 0 ) y = 1 0 4 p s = [ i p . g e t P i x e l ( x , y ) f o r x i n x s ] p r i n t p s
19. VARIOUS THRESHOLDING METHODS Fixed threshold Automatic threshold calculated from the histogram of the image E.g. Otsu's method Automatic local threshold Using varied thresholds for different regions
20. RUNNING A PLUGIN FROM A SCRIPT 1. Get a parameter for running a plugin. 1. Plugins → Macros → Record... 2. Run a plugin you want to use. 3. Copy a recorded code from a Recorder window. 2. Use I J . r u n ( ) with the parameters obtained above. I J . r u n ( " A u t o L o c a l T h r e s h o l d " , " m e t h o d = M i d G r e y r a d i u s = 1 0 0 p a r a m e t e r _ 1 = 0 p a r a m e t e r _ 2 = 0 w h i t e " )
21. COMMON PATTERNS OF FIJI SCRIPTING FROM JYTHON TUTORIALS Obtaining the current image imp, imp2, and imp3 are all the same. Generating a random image Obtaining a histogram i m p = I J . g e t I m a g e ( ) # i m p i s a n o b j e c t o f I m a g e P l u s c l a s s i m p 2 = W i n d o w M a n a g e r . g e t C u r r e n t I m a g e ( ) c = W i n d o w M a n a g e r . g e t C u r r e n t I m a g e i m p 3 = c ( ) f r o m j a v a . u t i l i m p o r t R a n d o m i m p = I J . c r e a t e I m a g e ( " A R a n d o m I m a g e " , " 8 - b i t " , 5 1 2 , 5 1 2 , 1 ) R a n d o m ( ) . n e x t B y t e s ( i m p . g e t P r o c e s s o r ( ) . g e t P i x e l s ( ) ) i m p . s h o w ( ) i m p = I J . o p e n I m a g e ( " h t t p : / / r s b . i n f o . n i h . g o v / i j / i m a g e s / b l o b s . g i f " ) s t a t s = i m p . g e t S t a t i s t i c s ( ) p r i n t s t a t s . h i s t o g r a m
22. MORE ABOUT IMAGE STATISTICS (1): QUANTITIES g e t S t a t i s t i c s function returns a object. g e t S t a t i s t i c s function can take a parameter to specify which quantities to calculate. Join them with | . AREA, MEAN, MIN_MAX, MODE, etc. An ImageStatistics object has attributes you can take values from. area, mean, min, max, mode, etc. i m p = I J . g e t I m a g e ( ) s t a t s = i m p . g e t S t a t i s t i c s ( M e a s u r e m e n t s . M E A N | M e a s u r e m e n t s . M E D I A N | M e a s u r e m e n t s . A R E A ) p r i n t " m e a n : " , s t a t s . m e a n , " m e d i a n : " , s t a t s . m e d i a n , " a r e a : " , s t a t s . a r e a ImageStatistics Measurements
23. MORE ABOUT IMAGE STATISTICS (2): ROI Region of Interest (ROI) is a part of the image for which you process the pixel data. i m p = I J . g e t I m a g e ( ) # O b t a i n t h e c u r r e n t i m a g e a s a n I m a g e P l u s o b j e c t # C a s e 1 # S e t t h e R O I o f t h e i m a g e . i m p . s e t R o i ( 1 0 , 3 0 , 2 5 , 2 5 ) # x , y , w i d t h , h e i g h t s t a t s = i m p . g e t S t a t i s t i c s ( M e a s u r e m e n t s . M E A N | M e a s u r e m e n t s . A R E A | M e a s u r e m e n t s . C E N T R O I D ) p r i n t " m e a n : " , s t a t s . m e a n , " c e n t r o i d : " , ( s t a t s . x C e n t r o i d , s t a t s . y C e n t r o i d ) , " a r e a : " , s t a t s # C a s e 2 r a d i u s = 2 5 # G e t a n R O I w i t h a n o v a l s h a p e , t h e n s e t i t t o i m p . r o i = O v a l R o i ( i m p . w i d t h / 2 - r a d i u s , i m p . h e i g h t / 2 - r a d i u s , r a d i u s * 2 , r a d i u s * 2 ) i m p . s e t R o i ( r o i ) s t a t s = i m p . g e t S t a t i s t i c s ( M e a s u r e m e n t s . M E A N | M e a s u r e m e n t s . M E D I A N | M e a s u r e m e n t s . A R E A ) p r i n t " m e a n : " , s t a t s . m e a n , " m e d i a n : " , s t a t s . m e d i a n , " a r e a : " , s t a t s . a r e a
24. A FEW IMPORTANT THINGS IN PYTHON
25. VARIABLES Variables are used for keeping track of data in a program. You can name a variable with your favorite way.
26. CLASSES AND OBJECTS IN PYTHON i m p = I m a g e P l u s ( ' h t t p s : / / w w w . g o o g l e . c o m / i m a g e s / s r p r / l o g o 1 1 w . p n g ' ) i m p . s h o w ( ) I m a g e P l u s is a class, and i m p is an object Class is a set of data (called attributes, fields, instance variables, etc.) and functions (methods) to manipulate them. Class is a template for generating an object (or instance). Each object has its own data. It may take some time to understand this concept of Object Oriented Programming (OOP), but this is very useful for writing a better organized code. . (dot) is used for accessing data or a method of an object or a class
27. USING MORE IMAGEJ FUNCTIONS
28. TWO WAYS TO CALL IMAGEJ FUNCTIONS 1. ImageJ/Fiji Application Programming Interface (API) This API is originally for Java, but Jython can call those functions class class class , etc. 2. Running a plugin by I J . r u n ( ) API reference IJ ImagePlus ImageProcesssor
29. HOW TO READ THE API REFERENCE MANUAL ImageJ API is written in Java, so you need to translate it to Python when reading it. 'Static' methods belong to a class. Other methods belong to an object. Java has a fixed 'type' of a variable and parameter, but there is not in Python. You still have to pass arguments of the right types.
30. READING THE API REFERENCE: EXAMPLE Code for loading an AVI movie and get the number of slices. # o p e n ( ) i s a s t a t i c m e t h o d o f A V I _ R e a d e r c l a s s . # T h i s r e t u r n s a n I m a g e P l u s ' o b j e c t ' i m p = A V I _ R e a d e r . o p e n ( ' s o m e n i c e m o v i e . a v i ' , F a l s e ) # g e t N S l i c e s ( ) i s a n ' i n s t a n c e m e t h o d ' o f I m a g e P l u s c l a s s . # T h i s r e t u r n s a n i n t e g e r . p r i n t i m p . g e t N S l i c e s ( )
31. API REFERENCE: EXAMPLE o p e n is a 'static method', which is called for A V I _ R e a d e r class. g e t N S l i c e is an 'instance method', which is called for an object of I m a g e P l u s . AVI_Reader ImagePlus
32. REAL-WORLD EXAMPLES
33. SOME EXAMPLES OF IMAGE ANALYSIS IN LAB Radial profile analysis Nanodot spacing analysis Batch image processing
34. RADIAL PROFILE https://gist.github.com/hirokai/9575307 Define ROIs of concentric circles. Calculate mean intensity for each of them.
35. NANODOT SPACING ANALYSIS https://gist.github.com/hirokai/9573785 https://gist.github.com/hirokai/9574844 Threshold the image. Use "Image → Adjust → Auto Threshold... " command. Detect particles and collect coordinates. Use "Analyze Particle..." command. Use Delaunay triangulation algorithm to find nearst neighbors. Write some custom code in Python. Draw the result Use ImageJ APIs to draw lines.
36. BATCH PROCESSING Simple batch processing of multiple image files Process → Multiple Image Processor This applies the same script to all image files in a folder to generate result images.
37. MORE INVOLVED BATCH PROCESSING WITH SCRIPTING Strategy Define the part of script for image processing as a function Call that function multpiple times with varied parameters to process multiple files and conditions
38. BATCH PROCESSING: ANNOTATED EXAMPLES Minimum code for processing multiple images with different parameters. Uses a csv file as an input of parameters for processing. All parameters are written in a separate csv, and you don't have to modify the script for every run. https://gist.github.com/hirokai/9575273
39. ADVANCED TOPICS
40. INTERFACING FIJI/JYTHON WITH OTHER LANGUAGES AND SOFTWARES Many different softwares are used in scientific computing Matlab, Igor, ImageJ/Fiji, Python (numpy, scipy), etc. By combining those things, you can do even better jobs more easily. Various ways Interfacing by files. Generate data file from Fiji, and call an external program to process that file, or Generate the result as a file (plain text, csv, Excel, image, etc.) Interfacing by redirect. Interfacing by HTTP server/client
41. INTERFACING BY FILES The simplest and most versatile way.
42. WRITING BETTER PROGRAMS Use variables wisely Name variables in a concise yet meaningful way. Don't reuse the same variable for different purposes in one script. Extract the common operation into a separate function. There shouldn't be the duplicate of the same code in different parts of a program. Use functional programming (map, filter, reduce) to keep the code simple. Operation with "side effect" should be minimized. Use immutable variables instead of mutable ones. Collection of common bad situations in development. Don't Repeat Yourself (DRY) Anti-pattern
43. IMPLEMENTING PRIMITIVE IMAGE OPERATIONS "BY HAND" It's useful to better understand how they work. Thresholding Gaussian filter Median filter Edge detection
44. FILTERS Linear filters Gaussian filter Convolution filter Laplace filter Non-linear filters Mean filter Median filter Minimum/maximum filter
45. LINEAR FILTERS Applying a linear filter is multiplying a matrix (kernel) to a matrix of pix Gaussian blur, Linear filters have some nice properties Separable to two 1D filters Commutative: A * B = B * A Associative: ( A * B ) * C = A * ( B * C ) Linear: ( a A + b B ) * C = a A * C + b B * C http://www.cs.utexas.edu/~grauman/courses/378/slides/lecture
46. COMMON PITFALLS Be careful about overflow of pixel values. Image type Integer: 8-bit, 16-bit Float: 32-bit