Showing papers by "Sid-Ahmed Berrani published in 2014"

Arabic text detection in videos using neural and boosting-based approaches: Application to video indexing

Abstract: Text detection in videos is a primary step in any semantic-based video analysis systems. In this work, we propose and compare three machine learning-based methods for embedded Arabic text detection. These methods are able to detect Arabic text regions without any prior knowledge and without any pre-processing. The first method relies on a convolution neural network. The two other methods are based on a multi-exit asymmetric boosting cascade. The proposed methods have been extensively evaluated on a large database of Arabic TV channel videos. Experiments highlight a good detection rate of all methods even though neural network-based method outperforms the other ones in terms of recall/precision and computation time.

A Novel Modeling for Video Summarization Using Constraint Satisfaction Programming

Abstract: This paper focuses on automatic video summarization. We propose a novel modeling for summary creation using constraint satisfaction programming (CSP). The proposed modeling aims to provide the summarization method with more flexibility. It allows users to easily modify the expected summary depending on their preferences or the video type. Using this new modeling, constraints become easier to formulate. Moreover, the CSP solver explores more efficiently the search space. It provides more quickly better solutions. Our model is evaluated and compared with an existing modeling on tennis videos.

Boosting-based approaches for Arabic text detection in news videos

Abstract: In this paper, we propose two boosting-based approaches for Arabic embedded text detection in news videos. The first approach uses Multi-Block Local Binary Patterns features whereas the second one relies on Haar-like features. Both approaches learn text and non-text classes using a multiexit asymmetric boosting cascade. Bootstrap has also been used in order to improve the rejection ability of the classifiers. Text localization is performed by a sliding window search on a multiscale pyramid of the input image. The proposed approaches have been evaluated on a large database of images coming from 4 different Arabic TV channels.