A critical investigation of recall and precision as measures of retrieval system performance
ACM Transactions on Information Systems (TOIS)
Automatic Caption Localization in Compressed Video
IEEE Transactions on Pattern Analysis and Machine Intelligence
Automatic text detection and tracking in digital video
IEEE Transactions on Image Processing
Localizing and segmenting text in images and videos
IEEE Transactions on Circuits and Systems for Video Technology
A spatial-temporal approach for video caption detection and recognition
IEEE Transactions on Neural Networks
Hi-index | 0.00 |
Captions (or overlay texts) play an important role in video content understanding. In this paper, an algorithm is proposed to detect captions in MPEG compressed video. First, energy features, which are used to find candidate caption blocks, are extracted from DCT coefficients. Second, temporal information is employed to verify these candidate blocks. Then, a new region growing method named "density-based region growing" is proposed to connect these blocks into candidate text regions. Finally, the regions are identified as caption or non-caption by structural information of caption regions. Experiments are conducted on news videos and it is shown that the algorithm is feasible and effective in finding captions.