1. BLIND SEPARATION OF
IMAGE SOURCES VIA
ADAPTIVE
DICTIONARY LEARNING
PROJECT GUIDE SUBMITTED BY
Mr.R.Devaraj B.E., D.Mohan raj
D.Raja
A.Thangapandi

2. AREA OF PROJECT
• Digital Image Processing

3. SOFTWARE REQUIRED
• MATLAB 7.14

4. ABSTRACT
• we address this problem and attempt to give a
solution via fusing the dictionary learning into the
source separation
• We first define a cost function based on this idea and
propose an extension of the denoising method in the
work of Elad and Aharon to minimize it

5. INTRODUCTION
• A local dictionary is adaptively learned for each
source along with separation.
• To improve the quality of source separation even in
noisy situations

6. EXISTING SYSTEM
• SMICA (spectral matching Independent Component
Analysis)
• PCA (Principal Component Analysis)
• MMCA(Multichannel Morphological Component
Analysis)

7. DRAWBACKS
• to learn a specific dictionary for each source
from a set of exemplar images
• no prior knowledge about the underlying
sparsity domain of the sources

8. PROPOSED SYSTEM
• To adaptively learn the dictionaries from the mixed
images within the source separation process
• Hierarchical scheme such as the one in MMCA.
ADVANTAGES
• enhances the separability of the sources.

9. BLOCK DIAGRAM

10. FUTURE ENHANCEMENT
• we further reduce the time complexity by choose the
different size of the patches