Option B

Option B has 4 layers.

Layer 1B identify lines.

layer 2B divide images into 4 sections and create features for each section.

Layer 3B combine the 4 feature vectors into 1 feature vector.

Layer 4B identify the feature vector.


Layer 1B

We suggest you use the default setting.



Layer 2B

1. Delete Old Data
This step deletes the old problem. If you do not do this step, the new problem will be added to the olde problem.

2. Select a folder
This step select a folder to obtain features.

3. Specify Number of features and Layer size, say (width, height) = (50%, 50%);
This step define the number of feature, which should be between [200, 500].
layer 2B divide images into 4 sections; the default section size is (50%, 50%). We suggest you will this default setting for now.

4. Get Feature
This step gets feature vectors for each of the 4 sections. The results are in four folders.

5. Check Results: get rid of almost black, almost white images 
This step will visually represent the features. delete images that are almost black and almost white.


6. Move data
This step move the obtained feature to its final location.

7. Repeat the above steps until all of the data folders are processed
Now we have finished one folder of images. We will repeat the above 6 steps oever and over again untill all image folders are processed.

At Layer 2B, images are divided into 4 sections. At the of the layer 2 computations, features for each of 4 sections are defined as a results of going through many folders. The next layer will combine the features 







