SmartR T UCL Guillaume Lobet , Loïc Pagès and Xavier Draye A novel image analysis toolbox enabling quantitative analysis of root system architecture 6th of September 2012
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
SmartR T
UCL
Guillaume Lobet, Loïc Pagès and Xavier Draye
A novel image analysis toolbox enabling quantitative analysis
of root system architecture
6th of September 2012
1
What we do in our lab
2
What is SmartRoot?
• Semi-automated
• Root tracing software
• Morphology, topology
• Wide range of image
3
Why semi-automated?
4
Heterogenous image
4
Heterogenous image
5
Local information
5
Local information
6
Root system as a population
6
Root system as a population
7
Why semi-automated?
• Local information is needed
• Sample the root system
• Sample the image
8
How does it work?
9
Quick root tracing overview
9
Quick root tracing overview
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel valuepath
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel valuepath
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel valuepath
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel valuepath
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel value
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
path
pixel value
local threshold
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
path
local threshold
local threshold
path
pixel value
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
path
local threshold
local threshold
path
pixel value
local threshold
mouse click
10
Tracing algorithm - 1
node creation
search path
node creation
diameter check
append node
node centering
node centering
path
local threshold
local threshold
path
pixel value
local threshold
mouse click
11
Tracing algorithm - 2mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
11
Tracing algorithm - 2
Region of Interest
mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
11
Tracing algorithm - 2
pixel value
freq
uenc
y
MIN MAX
Region of Interest
mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
11
Tracing algorithm - 2
pixel value
freq
uenc
y
MIN MAX
Region of Interest
p
mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
11
Tracing algorithm - 2
pixel value
freq
uenc
y
MIN MAX
Region of Interest
THRESHOLD
p
T = p + MIN
2 + MAX
2
mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
12
Tracing algorithm - 3mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
13
Tracing algorithm - 4mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
14
Tracing algorithm - 5mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
14
Tracing algorithm - 5mouse click
node creation
search path
node creation
diameter check
append node
node centering
node centering
local threshold
local threshold
Diameter freeze
15
Object oriented
Nodes as individual objects
Root as individual object
15
Object oriented
Nodes as individual objects
Root as individual object
Individual actions[move, crop, append]
15
Object oriented
Nodes as individual objects Topology
Root as individual object
Individual actions[move, crop, append]
16
Vector based
16
Vector based
Coordinates in the pixel space
[x, y]
16
Vector based
Coordinates in the pixel space
Vector coordinates
[x, y]
[root, pos]
17
Working withreal images?
18
Real life examplesLupin proteoid root
Potato root growth
Maize root architecture
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
primary root tracing
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
primary root tracing
identify proteoid LR of interest across images
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
primary root tracing
identify proteoid LR of interest across images
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
primary root tracing
identify proteoid LR of interest across images
annotate cluster position
19
Lupin proteoid rootsaeroponic, time lapse
[contrast + , resolution + ]
primary root tracing
identify proteoid LR of interest across images
annotate cluster position
spatial and temporal root development dynamics[ ]
19
Root growth in rhizotronsrhizotron, apex tracking
[contrast - , resolution + ]
20
Root growth in rhizotronsrhizotron, apex tracking
[contrast - , resolution + ]
20
Root growth in rhizotronsrhizotron, apex tracking
root tracing
[contrast - , resolution + ]
20
Root growth in rhizotronsrhizotron, apex tracking
root tracing
mark positioning
[contrast - , resolution + ]
20
Root growth in rhizotrons
[growth = f(type, age, position)]
rhizotron, apex tracking
root tracing
mark positioning
[contrast - , resolution + ]
20
Maize root architectureaeroponic
[contrast - , resolution - ]
21
Maize root architectureaeroponic
[contrast - , resolution - ]
trace the primary
21
Maize root architectureaeroponic
[contrast - , resolution - ]
trace the primary
automatic lateral tracing
21
Maize root architectureaeroponic
[contrast - , resolution - ]
trace the primary
automatic lateral tracing
diameters, root length interbranch, angles [ ]
21
Take home message- Morphology, topology
- Wide range of: • image quality• plant species• biological questions
- Vector based and object-oriented
- Annotation tool
- User interface, data export, ....
22www.uclouvain.be/smartroot
Acknowledgments
- Tristan Lavigne- Geoffrey Berguet- Hélène Cordonnier- Chantal Lemarié
Beta testers
Funding
23
Related Documents
