Imagebased Evaluation of Videoacquired Research Skills

1
Image-based Evaluation of Video-acquired Research
Skills

• Unmil Karadkar, Marlo Nordt
• Richard Furuta

Cody Lee Christopher Quick
Center for the Study of Digital Libraries The
Department of Computer Science
Michael E. DeBakey Institute Veterinary
Physiology Pharmacology
Texas AM University
2
What do these images have in common?

• Biomedical Images
• Cognitive Perceptual Tasks
• Expertise Matters

3
Our Domain Bat Lab

• Cardiovascular Research
• Related to the blood circulatory system
• Observe bat wings under a microscope
• Pallid Bats have transparent wings
• Non-invasive, in-vivo studies
• Normal and modified conditions

4
Basic Research Skills

• Artery and Vein recognition
• Size differentiation
• Vein, Venule, Capillary
• Lymphatic vessel identification

5
Our Question
How do we know what
training is most effective?

Testing
And because were CSDL
Does the interface matter when testing? Why might
it matter?
Layouts differ in context clues
6
Image Layout Contexts
7
Experimental Design 1

• 3 Layouts
• Thumbnail
• Scrolling
• Montage
• 16 images per set

• 3 Tasks
• Artery/Vein recognition
• Size Estimation
• Lymphatic vessel wall identification

8
Experimental Design 2

• Balanced across subjects

Round 1
Round 2
Round 3
Version
9
Experimental Design 3

• PowerPoint for presenting layouts
• Feedback
• Verbal
• PowerPoints Pen feature
• Validation of answers
• Panel of experts
• Accurate measurements

10
Subjects

• Trained researchers in the bat lab
• 9 experts
• Research experience 2 semesters
• 6 novices
• Research experience 1 semester only
• Age group 18 to 35
• 4 female, 11 male

11
Artery/Vein Recognition Results
(p0.005)
Learning (p0.0004)
No Learning
12
Size Estimation Results
Experts performed better than novices (plt0.002)
General understanding of the order of blood
vessels does not translate into accurate
estimation of size
13
Lymphatic Identification Results
Experts worked best with Montage (p0.035)
Experts and Novices use different strategies for
applying contextual information
14
Future Work

• Explore other image layouts
• Static and dynamic collages
• Train subjects on new interfaces
• Improve learning time
• Increase productivity
• Investigate mechanisms for conveying size of
  objects within images

15
Further Information
http//ebat.tamu.edu Center for the Study of
Digital Libraries Michael E. DeBakey Institute