Interactive retrieval of video using pre-computed shot-shot similarities

Abstract

A probabilistic framework for content-based interactive video retrieval is described. The developed indexing of video fragments is originated from the probability of the user's positive judgment about key-frames of video shots. Initial estimates of the probabilities are obtained from low-level feature representation. Only statistically significant estimate are picked out, the rest is replaced by an appropriate constant allowing efficient access at search time without loss of the search quality and leading to improvement in most experiments. With the time, these probability estimates are updated from the relevance judgment of users performing searches, resulting in further substantial increase in mean average precision.