JPEG metadata processor

import fnmatch
import os
import pyexiv2

matches = []
dict1 = {}

dict1 isn't a great name because its hard to guess what it might be for.

# The aim of this script is to recursively search across a directory for all
# JPEG files. Each time a JPEG image is detected, the script used the PYEXIV2
# module to extract all EXIF, IPTC and XMP data from the image. Once extracted
# the key (ie. "camera make" is generated and it's respective value
# (ie. Canon) is then added as the value in a dictionary.
for root, dirnames, filenames in os.walk('C:\Users\Amy\Desktop'):

Your code will be a bit faster if you put stuff into a function rather than at the module level.

  for filename in fnmatch.filter(filenames, '*.jpg'):
      matches.append(os.path.join(root, filename))
      for entry in matches:

ok, you add each file you find to the matches. But then each time you go over all the matches. I doubt you meant to do that. As its stand, you are going to look at the same file over and over again

          metadata = pyexiv2.ImageMetadata(entry)
          metadata.read()
          x = metadata.exif_keys
          y = metadata.iptc_keys
          z = metadata.xmp_keys

x y and z aren't great variable names. But its seems you aren't really interested in them seperately, just in all the keys. In that case do

keys = metadata.exif_keys + metadata.iptc_keys + metadata.xmp_keys

to combine them all into one list. (I'm guessing as to what exactly these keys are but, I think I it'll work.)

          for key in x:
              value = metadata[key]
              dict1[key] = value.raw_value

I'd combine the two lines

dict1[key] = metadata[key].raw_value


          for key in y:
              value = metadata[key]
              dict1[key] = value.raw_value
          for key in z:
              value = metadata[key]
              dict1[key] = value.raw_value
              print dict1

If you put combine x,y,z as above, you'll only need the one loop