Bitan:imageJ

From OpenWetWare

Jump to: navigation, search

ImageJ Protocol at 4x Aida Attar (Dec 2013)


1. File > open

2. Image > Type > 8-Bit

3. Process > Subtract background a. Rolling ball radius: 10-100, play with it b. Check “Light background” is ticked

4. Process > Enhance contrast a. 0.5-1% but doesn’t seem to affect actual number, just makes image nicer

5. Process > smooth (doesn’t do much sometimes)

6. Image > adjust > Threshold a. Move Threshold to appropriate (Usually higher than what they have it), Select “Set” then “OK” b. Select paintbrush tool, this enable you to erase areas. (double click on the paintbrush tool allow you to change the size of the paintbrush and the color) can change color by using eyedrop tool.

7. Set scale once (every time open ImageJ) see bullets for steps below

8. Select region of interest with Freehand selection tool

9. Analyze > measure (to get size of region selected) a. Record in excel this “area selected” measure for each section) b. Record after each image because imageJ tends to crash

10. Analyse > Analyse particles a. The values will differ depending on what kind of staining. Play around for optimum b. Size (pixel^2): 0- infinity (or 15-infinity to exlude single dots) i. if you measure in pixels or your picture has a lot of dots in the background, will have to play with this number

c. Circularity: 0-1 i. Might need to play with this number, maybe .1-.9 if circular objects would obviously be artifacts.

d. Show: masks. Use this to ensure the area stained are analyzed. e. Check: Display results, summarise, exclude on edges, in situ show. f. Record in excel with these titles “count, total area stained, avg stain size, stain % of total area”


• Set the Scale. By o Select line drawing tool; draw a line across (horizontally) the photo. • To ensure the line is parallel to bottom of the photo, draw the line while looking at the ImageJ toolbox. The angle should read 0 and the length 2560. • Top left corner of the photo should say 2560x 1920 pixels; 8-bits, 4.7MB o From Tanya’s protocol, at 10x magnification, “200pixels = 96.15um, so 2560pixels would = 1230.072um” o So at 4x, 2560pixels would = 3075.18 o Whole box size • 2560x1920 pixels = 4,915,200 pixels squared • 492.03x369.02 micrometer = 181,570.141 um squared • 27.07121 pixels^2 per 1 um^2 • Analyse > set scale o Distance in pixel should be at 2560 if you drew the straight line completely across. o Known distance = 492.0288um o Units of length = micrometer o Select global, this enable the same scale when you open the next photo. However the scale will only stay the same for this time ImageJ is in use. E.g The scale will need to be reset if you are processing more photos tomorrow as ImageJ quit overnight.

Personal tools