Extraction of Image features

Thank you for reading my question.

I'd like to detect the interesting structure (synapses of neurons) in the images. I have written the code to find the candidate ROI (range of interest) which may contain a synapse, and I plant to write the program to judge whether the ROI truly contains synapse or not.

So, the extraction of the features vector from the candidate ROI is needed. Also, to make the judge precise, I was told that the number of dimensions of this vector should be over 10,000.

I am searching for the code source, but I can't find it.

I appreciate it if you would advise me about this matter.

[Walter Roberson]

Could you link in a couple of sample images?

[Kotaro]

Thank you for commenting. I am truly sorry that I didn't notice your comment.

Could you look at fig.3 on the following link? http://dsresearch.stanford.edu/community/archive_issue_03.html

There are some characteristic features. For example, synaptic vesicles is mark for synapses when human find synapse from raw EM image. Also, PSD (post synaptic density) is quite high in electronic density ,so this looks black bar.