Image Captioner

In a nutshell the project would create a software that could automate the production of music videos. User of the software would provide the software with a song in a digital form and the program would help the user to annotate the song with tags. This could be done either manually by the user or with the help of speech recognition technology. After the software has knowledge of what is the song about, it would go out and fetch pictures that match the keywords of the song. The software would place those photos in linear order to form a music video which combines the music and a changing stream of photos. The source for the photos could the users own collection or photos that can be used with proper creative commons licenses. The software would also create end credits and licensing information. Here is an example of what it could produce: http://video.google.com/videoplay?docid=8060206257543341917