Presentation for the current results on red blood cell study Jia-Rong, Yeh March 6, 2009.

Presentation for the current results on red blood cell study

Jia-Rong, Yeh

March 6, 2009

Contents

Aims of this study Data format Movie player Mode function Spectral analysis Linguistic analysis

Aim of this study

Health of red blood cell (RBC) affects the results of blood transfusions.

The fluctuation on surfaces of RBC performs nonlinearly dynamical property.

As a pilot study, we try to quantify the dynamical property of RBC as a factor for aging of RBC.

In this study, many different analysis techniques have been used to quantify the dynamic property of RBC

Saving movies using a new data format The original data was saved pixel by pixel

and frame by frame. I developed a new data format to store a

movie with a small requirement of space. Space using the original format is 800 MB,

but only 130 MB using the new data format for a movie.

The new format

The boundary map of RBC (matrix img_bund)Grey area presents the pixels outside the RBCThe values of white are is the corresponding column number in matrix ‘img’. The red pixel is pixel mapped to the column 1 of img.

Pixel 1 2 3 nThe matrix of img. Each row presents a frame of movie. Each column vector presents the time series of a pixel. The red column presents the pixel 1 shown (2,3) in img_bund.

0 0 0 ………. 0

0 0 1 ……….. 0

0 4 5 ……….. 0

:::

:::

:::

0 0 0 0

Frame 1 ** ** ** ……………. **

Frame 2 ** ** ** ……………. **

Frame 3 ** ** ** ……………. **

:::

:::

:::

:::

Frame 5000 ** ** ** ……………. **

Movie player

The data can be saved in one fourth of the original space using the new data format

Movie player is a Matlab program used to convert the 5000 row vectors to a sequence of frames as a movie.

Original movies of new and old RBC

Mode decomposition using EMD

Before mode decomposition, the resolution of movie had been reduced to save calculation time.

Here, a frame with resolution around 200 × 200 had been reduced to 65 ×65

Then, the time series of each pixel inside the boundary of RBC had been decomposed into the first 7 IMFs

Movie player for mode functions

Seven mode functions can be assembled using the IMFs of pixels inside RBC.

A comparison between mode functions of old and new RBCs was shown using the first 4 mode functions of both.

Movies of the first 4 mode functions for both new and old RBC

Spectral analysis

Based on the characteristics of IMF, energy spectrum of each IMF should concentrate on a frequency band.

Therefore, I used 3-D plots to display the energy spectra for the pixels inside the boundary of RBC.

The pixels are sorted by the total energy of fluctuation.

3-D Spectrum of two RBC

New RBC Old RBC

Linguistic analysis (Preparation)

Madalena gave me 26 RBCs ( 13 new and 13 old) with MSE, alfa, and SD maps

I tried to figure out the connection among the 26 RBC and the original data sources saved in directories of \RBC\original\new, \RBC\original\old, and \RBC\original\Sep-2-7

I found the original data sources saved in \RBC\original for 20 (13 new and 7 old) out from the 26 RBCs used in previous publications

Display three maps of RBC

New RBC Old RBC

Application 1 of Linguistic analysis

Divide the area inside RBC into three regions: core, core border, and ring

Sort the pixels to three groups according to the region.

Use the ensemble ranks and probabilities of words via the pixels locate on core region as the baseline to plot distance map.

Distance map

New RBC Old RBC

Application 2 of Linguistic analysis

Plot the distributions of entropy for old and new RBC

Define the ranges of both high and low entropy for old and new RBC

Randomly select 300 pixels for both high- and low-entropy groups from 3 old and 3 new RBC

Conduct statistical analysis among groups

Entropy distributions for new and old RBC

6 7 8 9 10 11 12 130

50

100

150

200

250

300

350

400

Complexity (MSE)

Num

ber

of o

ccur

ence

New RBC 3

6 7 8 9 10 11 12 130

20

40

60

80

100

120

Complexity (MSE)

Num

ber

of o

ccur

ence

Old RBC 2

New RBC Old RBC

4 5 6 7 8 9 10 11 12 130

50

100

150

200

250

300

350

Complexity (MSE)

Num

ber

of o

ccur

ence

New RBC 6

4 5 6 7 8 9 10 11 12 130

20

40

60

80

100

120

140

160

Complexity (MSE)

Num

ber

of o

ccur

ence

Old RBC 7

The following analysis is being continued ……….

In fact, temporary results of linguistic analysis are not good enough to show the difference between new and old RBC.

Analysis should be continued ……..

Conclusions

Different techniques have been used to analyze the RBC data

The original data sources is messy, it is time-consumption to rearrange the database.

It is necessary to design parameters for quantifying the characteristics of mode functions.