Download v AB - cse.sc.edu

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Mitosis wikipedia , lookup

List of types of proteins wikipedia , lookup

Organ-on-a-chip wikipedia , lookup

Amitosis wikipedia , lookup

Transcript
A Cell Image Segmentation
Algorithm By Simulating
Particle Movement
Project report of Computer Vision
Xijiang Miao
Outline
Introduction
 Related works
 The algorithm
 Potential problems

2
Cells under microscope
mitosis
Gap
synthesis
apoptosi
s
3
Mission: Tell apart each cells

Knowing the number of cell is helpful
 Extract
RNA, …
 Currently, the number of cells is manually
counted.

Classifying cells in different phase is
valuable.
 Check
the effect a treatment.
 Integrate into cell sorting machine.
4
Revisit the image
5
-0.05
-0.10
-0.15
-0.20
Intensity
0.00
0.05
0.10
Voting Based Algorithm
-0.2
-0.1
0.0
X axis of image
0.1
0.2
6
Experimental Result of Simple
Voting
7
A recent published vote based
algorithm
Yang, Q. et al, Perceptual Organization of Radial Symmetries,
Proceedings of (CVPR’04)
8
Watershed algorithm
Fig. 2. Building dams at the places where
the water coming from two different
minima would merge.
Vincent, L. and Soille, P. Watersheds in Digital Spaces: An Efficient Algorithm
Based on Immersion Simulations. IEEE TRANSACTIONS ON PATTERN ANALYSIS
AND MACHINE INTELLIGENCE, VOL. 13, NO. 6, JUNE 1991
9
Watershed…
10
Watershed in ImageJ
ImageJ: http://rsb.info.nih.gov/ij/
Watershed plugin: Biomedical Imaging Group
http://bigwww.epfl.ch/sage/soft/watershed/
11
Think the pixels as particles
A



AB
Think each pixel is a particle with
its mass and velocity.
mAB = mA + mB
conservation of momentum



B
 mAvA + mBvB = (mA+mB)vAB
 vAB = (mAvA + mBvB)/(mA+mB)
Interpretation of mass and
velocity
12
Think the pixels as particles (2)

Average Mass and momentum


Weighted by their mass.
The overall goal is to
1.
2.
bring down the effect of noise and
accelerate the process.
13
The algorithm
1.
2.
Initialize the mass and speed.
Repeat
1.
2.
3.
4.
3.
4.
Move particles at their speed and direction
Once two particles collide together, merge these two
particles and recalculate their speed and mass.
Adjust the speed and mass according to its neighbors.
Record their paths
Until some terminate condition
Segment the image according to paths
14
Experiment result
15
Parameters and Options

Initial parameters

Mass


Speed



Gradient w/ tangent direction
Markers
Terminate condition




Gradient + ?
Limited Steps
Sand-box
Compete
sigma
16
Another example shows some
problems
17
The result
18
Problems and workaround

Global color changes
 Normalize

the marginal distribution.
Big blank area
 Use
different initial mass value
19
Question/
Suggestion
20