Imaging Ethics - Responsible Conduct in Biomedical Research

Imaging Ethics - Responsible Conduct in Biomedical Research

Loading
Loading Social Plug-ins...
Language: English
Save to myLibrary Post page as Slide Download PDF
Go to Page # Page of 51

Description: This medical presentation is about biomedical research responsibility and Imaging ethics. Learning Objectives for this Lecture to understand the differences between image file formats. Learn about issues involved with preparing images for presentation.

Understand what the “image size” box in Photoshop means and how it should be used. Learn what digital image manipulations are really doing to the image. Understand which types of image processing are allowable.

Have a comprehensive understanding of the ethics of image processing. Some computer programs used for image processing employ lossy compression like Adobe Acrobat may be default configured to apply JPEG compression to any images embedded in the document being converted to PDF. Microsoft PowerPoint should not be used for compiling image montages for publication.

The scalebar on the image will be carried through all processing and resizing required, Resizing of an image will render a magnification listed in the figure legend incorrect. Do not use Cut/Paste for inserting images into a document – process results in decreased image resolution. When it comes to figures and images, we all want those many hours of work to be represented in as good an image as possible.

But beware: a 'picture-perfect' image could land you in trouble, no matter how innocent your intentions. After some consultation with experts, the Nature family of journals has come up with a guide for how to handle digital images. In brief, the less you do, th ...Please navigate Paper pages for more details.

 
Author: Douglas J. Taatjes PhD (Fellow) | Visits: 1764 | Page Views: 2352
Domain:  Medicine Category: Imaging Subcategory: Professional Development 
Upload Date:
Link Back:
Short URL: http://medical.wesrch.com/pdfME1AU7CWXJOJC

Loading
Loading...



px *        px *

* Default width and height in pixels. Change it to your required dimensions.

 
Contents:
Imaging Ethics
Being a Scientist: Responsible Conduct in Biomedical
Research (ANNB 327)

Douglas J. Taatjes, Ph.D.
Department of Pathology and
Microscopy Imaging Center
College of Medicine
University of Vermont
February 6, 2014

Learning Objectives for this Lecture
 1.
 2.
 3.
 4.
 5.
 6.

Understand the differences between image file formats.
Learn about issues involved with preparing images for
presentation.
Understand what the “image size” box in Photoshop
means and how it should be used.
Learn what digital image manipulations are really doing to
the image.
Understand which types of image processing are
allowable.
Have a comprehensive understanding of the ethics of
image processing.

Considerations for Image Presentation

fred2blue.wordpress.com

Image File Formats
 TIF(F): Tagged Image File Format
 Highest resolution
 Loss-less (Lempel-Ziv-Welch compression; not widely supported)
 Supports wide range of bit depths
 Should always be used for image analysis

 JPEG: Joint Photographic Expert Group (.jpg)
 Most common form of lossy file compression
 Not suitable for scientific images for analysis
– Can alter dimensions, positions, intensities and colors
 Cosine transform (DCT)used changes the XY resolution of the

image and the intensity value of any given pixel
– JPEG2000 uses wavelet transform
– Slow save/conversion time
– Has not been widely implemented in scientific imaging
 Can be used for adding images to PowerPoint presentations

 PNG: Portable Network Graphic
 Lossless with high resolution

Image Format and File Size

24 Bit Color TIF = 5,631 KB

JPEG = 211 KB

Use JPEG files for PowerPoint presentations, but not for:
-Publications
-Quantitative evaluation

Image Format and File Size

TIF = 3893 KB

JPEG = 316 KB

PNG = 1710 KB

Degrees of JPEG Compression
TIF
8762 KB

80% JPEG
2745 KB

30% JPEG
627 KB

5% JPEG
167 KB

Does Image Conversion to JPEG File
Format Result in Loss of Data?

Does Image Conversion to JPEG File
Format Result in Loss of Data?
TIF Image

-

JPEG Image

Altered Pixels

Confocal Images, Optimal resolution (1152 X 1152 pixels)

Measures of Image size
 Pixel count
 512 x 512; 1024 x 1024; etc
 Determined during original image capture
 Physical size
 3.5 in x 3.5 in
 Resolution
 PPI – pixels per inch
Screen resolution – 96 ppi

Preparing Images for Publication
 Some computer programs used for image processing

employ lossy compression
 Adobe Acrobat may be default configured to apply JPEG

compression to any images embedded in the document
being converted to PDF (typically 72-150 DPI)
 Always create PDF using “Print” function, not from Adobe toolbar
 Set Conversion Setting to “Press Quality”
 Disable any image compression settings (there may be many)

 Microsoft PowerPoint should not be used for compiling

image montages for publication
 Resizing images can result in loss of data
 Compressing PPT file likely applies JPEG compression to embedded

images
 Was designed to prepare presentations at 72 dpi screen resolution
– TIFF output files will be at 72 dpi (even for intended printing)
– Print resolutions: 300 dpi (images); 1200 dpi (text)
 Many journals will not accept images assembled in PowerPoint

Preparing Images for Publication
(continued)
 Always include a scalebar on the original image
 Calibration of your microscope should be

established for each objective lens
 The scalebar on the image will be carried
through all processing and resizing required
 Resizing of an image will render a magnification
listed in the figure legend incorrect
 Do not use Cut/Paste for inserting images into a
document – process results in decreased image
resolution

Stage Micrometer to Calibrate
Microscope Magnification

4X objective lens
20 X objective lens
Graticules Ltd., 2 mm/10 µm per division

RGB Versus CMYK Color Space:
Why Do My Color Images Look Duller on the
Printed Page Than on My Monitor?
 RGB represents additive





color
CMYK represents
subtractive color
Possible colors formed
from RGB >> colors
formed from CMYK
Saturated colors not well
represented by CMYK
RGB colors produced from
back-lit device (i.e.
computer monitor), appear
more brilliant than those
from reflected pigments on
paper (CMYK)

http://www.m-graphix.com/featured_articles/photos/cmyk-rgb.jpg

Color Calibration Microscope Slide

Computer Monitor Calibration

Image Size Feature in Photoshop
I. Image from unknown device (confocal
microscope) defaults to 72 ppi

Image Size Feature in Photoshop
II. Image from unknown device can be resized dependent
upon input resolution

Adobe Photoshop CS5

Image Size Feature in Photoshop
III. Image from known device (in this case twain driver from
a flat-bed scanner) imports with correct resolution
displayed

Ethics in Digital Imaging
“Whenever you do a thing, act as if all the world
were watching."
Thomas Jefferson

http://richmondthenandnow.com

Journals Spreading the Word …
Special Seminar
EMMA HILL
Executive Editor
The Journal of Cell Biology, Rockefeller University Press
A PICTURE IS WORTH A THOUSAND WORDS:
IMAGE SCREENING AT THE JCB
OR
THE TEMPTATION OF IMAGE MANIPULATION
Monday, Oct. 5, 2009
12:15pm
Cannon Room – HMS
Host: Jennifer Waters

“The less you do,
the better…”
“When it comes to figures and images, we all
want those many hours of work to be represented
in as good an image as possible. But beware:
a 'picture-perfect' image could land you in trouble,
no matter how innocent your intensions. After
some consultation with experts, the Nature family
of journals has come up with a guide for how to
handle digital images. In brief, the less you do,
the better. Some image processing is acceptable,
but it must be applied equally to the entire image,
and the image-acquisition tools and imageprocessing software packages must be clearly
indicated. The use of any touch-up tools to
manipulate the image is not allowed.
All our images and figures are examined by our
production editors, but only those that raise red
flags (and I am being purposefully vague here)
are scrutinized more closely for image
manipulation. When such cases are identified,
we contact the authors for clarification and
explanation to resolve any potential problems
before publication. “

“Where the Deer and the Antelope
Play…”

Job

From: Nature 451:1034 (2008) “Acclaimed photo was faked.”

Altering of Missile Test Firing Image in
Photoshop

National pride and show of strength

BP: “Busy Photoshopping”
“Faked”Image
Shown by BP

Public relations

Actual Image

http://gawker.com/5591246/bps-photoshopped-command-center-picture

“Honest Abe” but not Honest
Photography

This portrait of President Abraham Lincoln
Taken around 1860 is a composite of
Lincoln's head and the Southern politician
John Calhoun's body. The first
permanent photographic image was
created in 1826 and the Eastman Dry
Plate Company (later to become
Eastman Kodak) was created in 1881.

http://www.cs.dartmouth.edu/farid/research/digitaltampering

A General Misrepresentation …

MATTHEW BRADY / LIBRARY OF CONGRESS

Civil War Generals, c. 1865
http://www.time.com/time/printout/0,8816,1927227,00.html
“Generally regarded as the world's first commercially successful photojournalist, Matthew Brady was also one of the medium's most
accomplished manipulators. In this group portrait of William Tecumseh Sherman and his top officers, he added one figure. For the
record, the men are, standing, from left: Oliver Otis Howard, William Babcock Hazen, Jefferson Columbus Davis
andJoseph Anthony Mower; seated, from left: John Alexander Logan, Sherman, Henry Warner Slocum and Francis P. Blair.”

Ernst Haeckel – “Ontogeny
Recapitulates Phylogeny”
Dog

Chick

Turtle

The first published concerns came from Ludwig Rütimeyer, a professor of zoology and comparative
anatomy at the University of Basel who had placed fossil mammals in an evolutionary lineage early in
the 1860s…Though he made no suggestion that embryo illustrations should be directly based on
specimens, to him the subject demanded the utmost "scrupulosity and conscientiousness" and an
artist must "not arbitrarily model or generalize his originals for speculative purposes" which he
considered proved by comparison with works by other authors. In particular, "one and the same,
moreover incorrectly interpreted woodcut, is presented to the reader three times in a row and with
three different captions as [the] embryo of the dog, the chick, [and] the turtle."
http://en.wikipedia.org/wiki/Ernst_Haeckel
http://bungabungabulan.tumblr.com/post/12024039009/ontogeny-recapitulates-phylogeny-ernst-haeckel

JENNIFER LAWRENCE ON HER DIOR
ADS: 'OF COURSE IT'S PHOTOSHOP;
PEOPLE DON'T LOOK LIKE THAT'

http://www.hollywoodreporter.com/fash-track/dior-premieres-jennifer-lawrence-as-423510

http://3.bp.blogspot.com/_mmBw3uzPnJI/SO9TtppvgQI/AAAAAAAAUQo/BPQaEo9_yhU/s1600-h/Before-and-after-Photoshop+14.jpg

Manipulation of Digital Images
 The Office of Research Integrity reports that

68% of its 2007-2008 cases involved
accusations of image fraud, compared with
12% a decade earlier
 Many journals now routinely scrutinize
images for possible unethical manipulation
 JCB reports that in 20-25% of accepted articles, at least one figure

required redoing due to failure to comply with instructions to authors

 Blogosphere has created “image sleuths”,

searching published images for
RetractionWatch
“irregularities”
NIH ORI

Human Embryonic
Stem Cells Derived
by Somatic Cell Nuclear
Transfer

Figure S5
Figure 6

Figure 2

Cell 153, 1228-1238; June 6, 2013

Figure 6

Image Falsification

Infection and Immunity 72(4):2280-2287, 2004 (Retracted October 2008)
B53
C95

Are There Any Issues with Presentation of
this Image?

“Detection of proteins by Quantum dots. Composite of six cultured fibroblasts (false-colored) from six different dishes immunolabeled
for a-tubulin and detected with Quantum dots analyzed by confocal microscopy: QD525 (blue), QD565 (green), QD585 (yellow),
QD605 (orange), QD655 (red) and QD705 (purple). Image: T.J. Deerinck and BNGG at NCMIR.”
B. Giepmans, Histochem Cell Biol 130:211 (2008)

A Digital Image is an Array of Pixel Values

homepages.inf.ed.ac.uk/rbf/CVDICT/CVFIG5/img8.png

http://huizen.ddsw.nl/bewoners/maan/imaging/camera/ccd_array1.gif

Would you Modify Data in a Spreadsheet?
Comparison of Lumen vs. Sinus
VWF
Vein
Lumen
Sinus
Difference
1
3,436,023 3,597,320
−301,346
2
606,896
70,446
−1,485,222
3
3,495,189 302,656
−2,018,080
4
3,155,665 54,556
−638,004
5
2,513,550 48,331
−202,980
6
4,034,410 173,815
−1,000,336
7
7,808,633 1,833,616
−3,130,730
8
2,492,265 1,128,727
−744,003
9
765,004
1,417,550
−765,667
10
2,913,875 738,397
−1,446,957
11
661,412
1,399,086
−1,243,851
Mean
1,919,856 −2,414,045
SEM
631,181
433,057
Paired t test p value
0.01

Difference

EPCR
Lumen

Sinus

Difference

−161,298

2,048,297

7,220,616

−5,172,319 1,051,538

1,352,884

536,450

1,003,169

2,744,583

−1,741,414 22,919

1,508,141

3,192,533

1,352,746

2,817,646

−1,464,901 807,512

2,825,592

3,101,110

2,489,943

4,454,583

−1,964,640 2,609,437

3,247,441

2,465,219

1,364,797

3,027,357

−1,662,560 520,135

723,115

3,860,595

1,509,630

2,526,787

−1,017,157 497,875

1,498,210

5,975,017

3,728,856

4,172,995

−444,139

885,543

4,016,273

1,363,538

43,080

2,390,000

−2,346,919 522,796

1,266,799

−652,546

31,534

3,367,644

−3,336,110 79,233

844,900

2,175,478

18,910

3,051,721

−3,032,811 292,957

1,739,914

−737,675

1,797,776

6,169,300

−4,371,524 0

1,243,851

−1,179,743
253,890
Subscribe
x