C# hashing and looking up images -

I am currently at the point where I can convert bitmap to byte arrays Suppose I have 26 images in one- With J, 26 byte arrays have been represented. Looking at an image, I want to use the byte array to see the correct letter instead of 26 comparison. Are there some ways to create byte arrays to prepare a hash code, which can be stored in the configuration file?

Alternatively, if there is a better (faster) approach than images of swans (assuming I have no access to underlying text representation) I would like to know a lot about them. For clarification purposes, assume that I have "a.bmp", "b.bmp" etc. Now I have an unknown image on the screen. I would have thought that image palm and a single lookup display would be the fastest way to identify positive. It should be faster than 26 different comparisons. If this assumption is wrong, then I appreciate the outline of optimal methodology.

Note: This is not the classic OCR problem (handwriting recognition etc.) because the letter will be presented equally every time. Therefore, the letter "a" will always produce the same hash code

You can get the C # algorithm An array of hash bytes again you can use the C # hash table datatype to map the hash to the character. However, you will still need to scan every byte of each bitmap, so there is Operation O (B * N) where B is the number of bytes in Bitmap and N is the number of characters. The size of specific bitmaps is not particularly efficient.

However, if this is OCR (optical character recognition) then this hash function will be completely useless. The value of hash is a pixel varies, even if specific optical noise from scanners or digital cameras prevents the two images of the same letter from hashing evenly. There are programsmatic OCR techniques, but this is a very deep subject and if you have an OCR problem then you may be better off using a built-in library.