A detection-based approach to broadcast news video story segmentation

Ma, Chengyuan; Byun, Byungki; Kim, Ilseo; Lee, Chin‐Hui

doi:10.1109/icassp.2009.4959994

Cited by 16 publications

(10 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Chaisorn et al [9] have studied the structure of news videos and noticed: Figure 1 exemplifies the structure of a typical news video. Also the ordering of news items may vary slightly from the broadcast station to station, but all have a similar structure and news categories [10], [11]. …”

Section: Major Breaking Newsmentioning

confidence: 99%

Anchor Person Detection using Haar-Like Feature Extraction from News Videos

Brindha¹,

Amsaveni²

2016

IJCA

View full text Add to dashboard Cite

The human face and facial feature extraction play a key role in person identification in the areas of video surveillance and access control on security reason. In this research work news video is taken for anchor person detection. Detecting anchor person from news videos give the time distribution to news readers and provide editorial support to journalist to find videos related to the particular person. First the video is converted into frames and the shot change detection algorithm is used to find scene changes and store the image in the database. Two different algorithms are used to find scene change detection such as color based shot detection and edge based shot detection. From these result, edge based shot detection performs well in finding shot changes more accurately. Second, segment the still image into skin region and non-skin region by using skin-color model based on its size and shape face region is identified. Third step, facial features like location of eye, nose and mouth are extracted to recognize face variations through haar-like features. It provides a possible ways to locate the positions of eyeballs, mouth centers, midpoints of nostrils and near and far corners of mouth from face image. This approach helps to extract features on human face automatically and improve the accuracy of face detection. Finally, anchor person is detected from news video by extracting SURF features of the given image. Experimental results show methods used in this research could locate facial features from face exactly and quickly.

show abstract

Section: Major Breaking Newsmentioning

confidence: 99%

Anchor Person Detection using Haar-Like Feature Extraction from News Videos

Brindha¹,

Amsaveni²

2016

IJCA

View full text Add to dashboard Cite

show abstract

“…Previous research shows that the presence of an anchor face is an important visual cue for story boundary detection [18], [30]. We first use an AdaBoost detector to detect human faces in video frames, and then use a regression classifier to discriminate anchor faces from other detected non-anchor faces.…”

Section: Video Featuresmentioning

confidence: 99%

Broadcast News Story Segmentation Using Conditional Random Fields and Multimodal Features

Wang

Xie

et al. 2012

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYIn this paper, we propose integration of multimodal features using conditional random fields (CRFs) for the segmentation of broadcast news stories. We study story boundary cues from lexical, audio and video modalities, where lexical features consist of lexical similarity, chain strength and overall cohesiveness; acoustic features involve pause duration, pitch, speaker change and audio event type; and visual features contain shot boundaries, anchor faces and news title captions. These features are extracted in a sequence of boundary candidate positions in the broadcast news. A linear-chain CRF is used to detect each candidate as boundary/non-boundary tags based on the multimodal features. Important interlabel relations and contextual feature information are effectively captured by the sequential learning framework of CRFs. Story segmentation experiments show that the CRF approach outperforms other popular classifiers, including decision trees (DTs), Bayesian networks (BNs), naive Bayesian classifiers (NBs), multilayer perception (MLP), support vector machines (SVMs) and maximum entropy (ME) classifiers.

show abstract

“…(iv) In Ma et al [10], a set of key events are first detected from multimedia signal sources, including a largescale concept ontology for images, text generated from automatic speech recognition systems, features extracted from audio track, and high-level video transcriptions. Then, a fusion scheme is investigated using the maximum figure-of-merit learning approach.…”

Section: 7mentioning

confidence: 99%

Automatic Story Segmentation for TV News Video Using Multiple Modalities

Dumont

Quénot

2012

International Journal of Digital Multimedia Broadcasting

View full text Add to dashboard Cite

While video content is often stored in rather large files or broadcasted in continuous streams, users are often interested in retrieving only a particular passage on a topic of interest to them. It is, therefore, necessary to split video documents or streams into shorter segments corresponding to appropriate retrieval units. We propose here a method for the automatic segmentation of TV news videos into stories. A-multiple-descriptor based segmentation approach is proposed. The selected multimodal features are complementary and give good insights about story boundaries. Once extracted, these features are expanded with a local temporal context and combined by an early fusion process. The story boundaries are then predicted using machine learning techniques. We investigate the system by experiments conducted using TRECVID 2003 data and protocol of the story boundary detection task, and we show that the proposed approach outperforms the state-of-the-art methods while requiring a very small amount of manual annotation.

show abstract

A detection-based approach to broadcast news video story segmentation

Cited by 16 publications

References 8 publications

Anchor Person Detection using Haar-Like Feature Extraction from News Videos

Anchor Person Detection using Haar-Like Feature Extraction from News Videos

Broadcast News Story Segmentation Using Conditional Random Fields and Multimodal Features

Automatic Story Segmentation for TV News Video Using Multiple Modalities

Contact Info

Product

Resources

About