On the Use of Standards for Microarray Lossless Image Compression

1
On the Use of Standards for Microarray Lossless
ImageCompression

• Author Armando J. Pinho, Antonio R. C.Paiva,
  and Antonio J. R. Neves
• Source IEEE TRANSACTIONS ON BIOMEDICAL
  ENGINEERING,VOL.53, NO. 3, MARCH 2006
• Speaker Ren-Li Shen

2
Outline

• Introduction
• Specialized Methods
• Standard Methods
• Experimental Results
• Sensitivity to Noise
• Conclusion

3
Introduction

• Standard image coding techniques applied to the
  lossless compression of microarray images
• JPEG2000
• JBIG
• JPEG-LS
• Try to overcome some of the drawbacks
• Image sources

4
Introduction

• Trying to identify compression technologies
• Provide efficient lossless compression results
• Offer relevant features for the microarray image
  compression problem

5
Outline

• Introduction
• Specialized Methods
• Standard Methods
• Experimental Results
• Sensitivity to Noise
• Conclusion

6
Specialized Methods

• Four published methods
• Jornsten et al.
• Gridding and segmentation
• Using a low complexity lossless compression
  algorithm
• SLOCO
• Hua et al.
• Transform-based coding technique
• Segmentation is using the Mann-Whitney algorithm
• Separately spots and background

7
Specialized Methods

• Faramarzpour et al.
• Locating and extracting the microarray spots
• Transforming the ROI(region of interest) into an
  one-dimensional signal
• Lonardi et al.
• Lossless and lossy compression algorithms for
  microarray images
• Fully automatic gridding procedure
• Similar to Faramarzpours method
• Split into two channels
• Foreground
• Background

8
Outline

• Introduction
• Specialized Methods
• Standard Methods
• Experimental Results
• Sensitivity to Noise
• Conclusion

9
Standard Methods

• JBIG
• Context-based arithmetic coding
• Focused on bi-level imagery
• JPEG-LS
• Predictive coding
• Lossless compression of continuous-tone images
• JPEG2000
• Transform based
• Providing a wide range of functionalities

10
Standard Methods(Experimental Results)

• Three different publicly available sources
• Apo AI set (32)
• ISREC set (14)
• MicroZip (3)
• Image size ranges from 1000 1000 to 5496 1956
  pixels

11
Standard Methods(Experimental Results)
12
Standard Methods(Sensitivity to Noise)

• 8bit-planes of cDNA microarray images are close
  to random and incompressible
• Result in some degradation in the compression
  performance
• Separated the images
• 8bit-planes
• 16bit-planes

13
Standard Methods(Sensitivity to Noise)
14
Outline

• Introduction
• Specialized Methods
• Standard Methods
• Experimental Results
• Sensitivity to Noise
• Conclusion

15
Conclusion

• JPEG-LS gives the best lossless compression
  performance
• JBIG was consistently better than JPEG2000
• The future of microarray image compression
  depends on special-purpose