histogram for visual letters

Hello,

Could someone help me with the following fuction :

I need to write a function that for an image IMG and a dictionary for visual letter , calculates the normalized histogram for the visual letters.

For each color pixel in the img image, the closest (in the Euclidean distance) visual word of C is identified and the frequency of the occurrence of the visual word in the associated histogram is incremented.

At the end, the histogram normalizes

Thank you

[dpb]

Homework Q????

[Image Analyst]

I have no idea what htis means. What is a "visual word of C"? What does each axis in your Euclidean space represent?

[Walter Roberson]

This could sort of make sense if every different "letter" was a different color.

But what I would expect is that there is a folder of test images, each of which holds an image of fixed size representing a letter, and there is a different folder of reference images, one per letter, and that you were intended to go through the folder of test images and find the euclidean difference between the test image and each of the reference images, and by finding the minimum, determine what the most likely match is among the reference images .