Learning Sequential Patterns for Lipreading
Learning Sequential Patterns for Lipreading
复制标题
学习唇读的顺序模式
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
R. Bowden
中科院分区：
文献类型：
--
作者：
Eng;R. Bowden
针对唇读问题，提出了一种基于S序列模式构建视觉序列分类器的机器学习算法(SP-Boosting)。我们证明了最优序列模式的穷举搜索是不可能的，因为巨大的搜索空间，并用一种新颖、高效的树搜索方法和一组剪枝准则来解决这个问题。重要的是，修剪策略增强了我们定位最佳序列模式的能力。此外，基于树的搜索方法考虑了训练集的增强权重分布。然后将这种时态S搜索方法集成到Boosting框架中，得到SP-Boosting算法。我们还提出了一种新的约束强分类器集，进一步提高了识别精度。通过在OuluVS数据库上执行多类识别，将得到的学习分类器应用于唇读。实验结果表明，该方法仅使用一小部分序列模式就能达到最好的识别效果。
This paper proposes a novel machine learning algorithm (SP-Boosting) to tackle the problem of lipreading by building visual sequence classifiers based on s equential patterns. We show that an exhaustive search of optimal sequential patterns is not possible due to the immense search space, and tackle this with a novel, efficient tr ee-search method with a set of pruning criteria. Crucially, the pruning strategies pres erve our ability to locate the optimal sequential pattern. Additionally, the tree-based searc h method accounts for the training set’s boosting weight distribution. This temporal s earch method is then integrated into the boosting framework resulting in the SP-Boosting algorithm. We also propose a novel constrained set of strong classifiers that fur ther improves recognition accuracy. The resulting learnt classifiers are applied to lipreading b y performing multi-class recognition on the OuluVS database. Experimental results show that our method achieves state of the art recognition performane, using only a small set of sequential patterns.