1. Introduction to Digital Image Analysis Part I: Digital Images
Kurt Thorn
NIC
UCSF
Image: Susanne Rafelski, Marshall lab

2. Many measurements of light intensity
What is a digital Image?
Many measurements of light intensity 1 3 5 4 2 6 13 20 11 14 44 75 81 45 12 28 98
255
234 78 27 94
215
194 68 18 39 66 63 35 8

3. Bit depth and dynamic range
# of bits
range 1 2 2 4 4 16 8
256
1-byte 12
4096
2-bytes 16
65536

4. Bit depth and dynamic range

5. Your monitor is an 8-bit display
Converting bit-depth
Your monitor is an 8-bit display
# of bits
range 1 2 2 4 4 16 8
256 12
4096 16
65536

6. Mapping values onto display
255
255
Display Intensity
Digital Number

7. Intensity scaling Computer screens are 8-bit
Publishers also want 8-bit files
255
Final Intensity
4095
Original Intensity
You lose information in this process –
values all end up as 255

8. Intensity scaling Max 255 Contrast Final Intensity Brightness 4095
4095
Original Intensity
Min

9. Brightness / Contrast adjustment

10. Be aware of how your software scales your image
Brightness/contrast
Be aware of how your software scales your image
Auto-Scale
Full Range

11. Be aware of how your software scales your image
Brightness/contrast
Be aware of how your software scales your image
High B/C
Full Range

12. Gamma correction Other contrast stretching transforms…. g=0.45
Output Intensity
g=2.2
g=1
Input Intensity
Other contrast stretching transforms….

13. Gamma adjustment
g=0.6
g=1
0.6
g=2.2

14. What are acceptable image manipulations?
JCB has the best guidelines
Brightness and contrast adjustments ok, so long as done over whole image and don’t obscure or eliminate background
Nonlinear adjustments (like gamma) must be disclosed
No cutting and pasting of regions within an image (e.g. individual cells)
Control and experiment must be treated identically

15. Lookup Tables (LUTs) 1 3

16. False coloring to bring out detail

17. False coloring to bring out detail

18. Lookup Tables (LUTs)

19. Lookup Tables (LUTs)

20. Color Images
Color images are made up of three gray scale images, one for each of red, green, and blue
Can be 8 or 16 bits per channel

21. Color Images

22. Color Images
0 0 0 = Or
255 93
134
214
137
249
209 89
255 90 39
185
139 93
231
185
255

23. Stacks: Sequences of images
Can represent time series(movies), z-positions, or other variables
813
790
827
860
873
771
807
882
921
1216
1062
940
815
962
1816
3422
2185
1140
803
1094
2395
3816
4134
1244
784
892
1464
1961
1689
1139
820
884
965
1082
1146
990
760
763
801
767
817
745
723
766
795
864
867
860
703
800
1016
1574
1299
849
744
829
1146
11117
2378
1030
756
792
908
1337
1149
883
757
874
880
823
831
651
790
915
1115
1086
871
799
1036
1851
2689
2300
1306
917
1688
2626
2354
2081
2304
983
2071
1790
4394
3187
2634
825
1671
2542
1820
1799
2206
803
1092
1823
2183
2314
1430 Z

24. File Formats Data sets can be big: 1392 × 1040 × 2 = 2.8MB
3-channels, 15 image z-stack, 200 time points:
2.8 × 3 × 15 × 200 = 25.2GB
Compression!

25. Compression Lossless vs. Lossy
Lossless: preserves all information in original image
Original image can be restored
Lossy: discards information not necessary for visual appearance
Original image cannot be restored
See:

26. File Formats Most portable: TIFF
8 or 16-bit, lossless, supports grayscale or RGB
Good: JPEG2000, custom formats (nd2, ids, zvi, lsm, etc.)
Lossless, supports full bitdepth
Custom formats often support multidimensional images
Not so portable
Generally Bad: Jpeg, GIF, BMP, etc.
Lossy and / or 8-bit

27. Software Tools Photoshop Gimp NIS-Elements AxioVision MetaMorph Zen
Acquisition + Analysis
Presentation
Photoshop
Gimp
NIS-Elements
AxioVision
MetaMorph
Zen
Slidebook
Micro-Manager
Analysis
Matlab
ImageJ
Imaris (3D visualization)
CellProfiler

28. Acknowledgements
Nico Stuurman