The Segmentation and Classification of Story Boundaries in News Video

Abstract

The segmentation and classification of news video into single-story semantic units is a challenging problem. This research proposes a two-level, multi-modal framework to tackle this problein. The video is analyzed at the shot and story unit (or scene) levels using a variety of features and techniques. At the shot level, we employ a Decision Tree to classify the shot into one of 13 pre-defined categories. At the scene level, we perform the HMM (Hidden Markov Models) analysis to eliminate shot classification errors and to locate story boundaries. We test the performance of our system using two days of news video obtained from the MediaCorp of Singapore. Our initial results indicate that we could achieve a high accuracy of over 95 % for shot classification. The use of HMM analysis helps to improve the accuracy of the shot classification and achieve over 89% accuracy on story segmentation.

Scene understanding for identifying persons in TV shows: Beyond face authentication

“…In the literature, there are several works that study scene analysis in sports videos, CCTV (Closed-Circuit Television) and news programs [9], [10], [11]. In [10], authors noticed 13 categories of shots in TV-news (interview, anchor, commercial, etc) and trained a decision tree to classify shots, then a Hidden Markov Model (HMM) to correct classification errors.…”

“…In [10], authors noticed 13 categories of shots in TV-news (interview, anchor, commercial, etc) and trained a decision tree to classify shots, then a Hidden Markov Model (HMM) to correct classification errors. Features used were shot durations, motion measurements, text and face detections.…”

Cross-lingual audio-to-text alignment for multimedia content management

“…The idea is that, since there are more silent frames in speech, the LSTER measure will be much higher for speech than that for music [21].…”

Story Boundary Detection in News Video using Global Rule Induction Technique

“…In order to alleviate the data sparseness problem, one way is to divide the framework into multi-layered tasks such as that employed in NLP research [9] that analyzes text documents at word, phrase and sentence levels. This multilayered approach was successfully adopted in [5,6,7].…”

“…It employed a hybrid approach and the HMM for the analysis at shot and story levels respectively. The system was evaluated on the TRECVID 2003 data set [3] and it achieved the best performance in the evaluations under story segmentation task [3,6]. But the disadvantages of the system are that it is computational expensive and not easy to scale up to larger and new corpuses.…”