Synchronization of lecture videos and electronic slides by video text analysis
MULTIMEDIA '03 Proceedings of the eleventh ACM international conference on Multimedia
High Performance Adaptive Fidelity Algorithms for Multi-Modality Optic Nerve Head Image Fusion
Journal of Signal Processing Systems
| Hi-index | 0.00 |
In this paper, we propose an approach for detecting signsfrom natural scenes. The approach efficiently embeds multi-resolution,adaptive search, and affine rectificationalgorithms in a hierarchical framework, with differentemphases at each layer. We combine multi-resolution andmulti-scale edge detection techniques to effectively detect textin different sizes. By using the cues from text inside theimage, we introduce affine rectification transformation torecover deformation of the text region caused by aninappropriate camera view angle. This procedure cansignificantly improve text detection rate and OCR (OpticalCharacter Recognition) accuracy. Experimental results havedemonstrated feasibility of the proposed algorithms. We haveapplied the proposed approach to a Chinese sign translationsystem, which can automatically detect Chinese text inputfrom a camera, recognize the text, and translate therecognized text into English or voice stream.