OK, my previous answer was a little 'confused'. Here is a link to a .NET 2.0 project (Visual Studio 2008 SLN format) that provides some basic 'extraction' functionality MetaExtractor ZIP (25Kb)
CODE SNIP:
// The Parser class extracts the data to hardcoded properties.
// it's 1200 lines - too many to post on StackOverflow
JpegParser parser = new JpegParser(path);
if (parser.ParseDocument())
{
Console.WriteLine("Parsed {0} {1}", System.IO.Path.GetFileName(path), parser.Title);
Console.WriteLine("Tags: {0}", parser.KeywordString);
Console.WriteLine("Description: {0}", parser.Description);
Console.WriteLine("Title: {0}", parser.Title);
Console.WriteLine("Rating: {0}", parser.Rating);
}
USAGE:
MetaExtractor "C:\Users\Craig\Pictures\anton-1.jpg"
OUTPUT:
== DeepZoomPublisher MetaExtractor v0.1 ==
Parsed anton-1.jpg Beach Photo
Tags: beach, blue sky
Description: Anton
Title: Beach Photo
Rating: 3
Press any key to exit...
Hope that helps more than my previous answer.