THE INTERPRETATION OF THE DIGITAL IMAGE BASED ON THE REGION

 

 

Figure 1． The Orginal Image

 

Figure 2. The Image of Region-based Segmentation

Figure 3. The Thresholding Image

Figure 4. Different Regions described by Regional Descriptors

 

The following figures are the several results of extracting from Figure 3.

 

Figure 5. The Patio Fence Region                            Figure 6. The Bricks Region

 

Figure 7. The Flowerpot Region 1                           Figure 8. Bricks Region 2 and Upper Fence

 

By applying this descriptor, in the image Figure 3, there are 178 connected components found and its Euler number is 52, which means there are 125 holes in the image. And the largest connected component is illustrated in Figure 4. The rest large connected component is illustrated in Figure 5-8. The following list is a basic statistical result among 178 connected components.

 

Regions

Number of Elements

Patio fense

7530

Bricks 1

6844

Bricks 2

2624

Fense 2

1182

Flowerpot 1

570

 

The next graphics are the results of applying the skeleton algorithm based on extracted regions.

Figure 9. Skeleton of Fence Region

 

Figure 10. Skeleton of Bricks Region

 

Conclusion

From the above processing, we can briefly draw a conclusion that the interpretation is regarded as a set of processes that are useful for gathering region attributes and helping to identify and understand the content of the image. And the interpretation processing lies in the segmentation techniques based on the region, especially thresholding and region-based segmentation etc. And there are different descriptors which are used to describe and interpret the image. In this case, the regional descriptor and the topological descriptor are applied for interpreting the digital image. Meanwhile, the exaction algorithm of connected components and skeleton algorithm are adopted for the image interpretation. So the strategy of  the image interpretation is simply viewed as two basic phases, 1) to segment the image into the ideal regions and 2) to take the appropriate methods to describe and interpret the content of the image based on the regions.