A hidden Markov model approach to the structure of documentaries

We have hand-segmented two very long documental-ies (100 minutes total) into their. component shots. As with other extended videos, shot distribution again appears to be log-normal. Shot lengths are similar to those in dramas, comedies, ol action films, but much shorter than those in home videos. The we of fades appeal-a to be an important device to signal transitions between semantic units. We have solcght evidence for shot compositon roles by means of Hidden Markov Models (HMMs). We find that camera motion (tilt, pan, zoom) is not significantly governed by rules. However, the bulk of the documentaries take the form of an alternation between commentators and several types of primary supporting material, additionally, the documental ies end with a visual summary. We find that the best approach is one that trains the HMM with labeled subsequences that have approximately equal elapsed time, rather than subsequences with an equal number of shots, or subsequences with shots aligned to some semantic event. This nlav reflects fundamental temporal limits oil human visual attention. We propose that such an underlying structure can suggest more human-sensitive designs for the analysis and graphic display of the contents of extended videos, for summarization browsing, and indexing.