中文说明：应用背景许多算法的存在，目的是解决稀疏表示字典学习问题。然而，没有全面的测试和基准测试，这些算法的存在，在与已知的字典稀疏表示的问题。主 ；这项工作的驱动力是一个工具箱，如字典学习 Sparco的缺乏；问题。认识到这样一个工具箱的社区的需要，我们设计出 ；smallbox-a MATLAB工具箱。关键技术SMALLbox是采用自适应稀疏信号处理的一种新的基础框架结构化表示。主要的SMALLbox是成为新的可证明的好方法探索的试验场和获得固有的数据稀疏模型，这是能够应付大规模和复杂的数据。研究的主要焦点稀疏表示的地区是开发具有可证明的性能可靠的算法和有界的复杂性。然而，这种方法根本不适用在许多情况下对于没有合适的稀疏模型是已知的。此外，稀疏模型的成功，在很大程度上取决于一个“字典”的选择，以反映自然结构的一类数据。从训练数据中推断字典是稀疏模型扩展的一个关键对于新的外来类型的数据。小箱提供评价这些方法的一个简单的方法在各种标准信号处理问题中对艺术的替代性。这是通过一个统一的接口，使三个无缝连接模块类型：问题，字典学习算法和稀疏求解器。此外，它提供了现有的国家的最先进的工具之间的互操作性。作为一个开源MATLAB工具箱的小盒子里，可以看到在稀疏的再现性研究工具表示研究社区。
English Description：
Background many algorithms exist to solve the problem of sparse representation dictionary learning. However, without comprehensive testing and benchmarking, these algorithms exist in the problem of sparse representation with known dictionaries. The main driving force for this work is the lack of a toolbox such as dictionary learning & nbsp; Sparco. Recognizing the needs of such a toolbox community, we designed & nbsp; smallbox-a matlab toolbox. The key technology, smallbox, is a new basic frame structured representation based on adaptive sparse signal processing. The main SMARTBOX is to become a new provable good method to explore the proving ground and obtain the inherent data sparse model, which is able to cope with large-scale and complex data. The main focus of sparse representation is to develop algorithms with proven performance and bounded complexity. However, this method is not applicable at all. In many cases, it is known that there is no suitable sparse model. In addition, the success of sparse model largely depends on the selection of a "dictionary" to reflect the natural structure of a class of data. Inferring dictionaries from training data is a key to sparse model expansion, especially for new foreign data. The small box provides a simple way to evaluate these methods and an alternative to art in various standard signal processing problems. It is through a unified interface that makes three seamless module types: problem, dictionary learning algorithm and sparse solver. In addition, it provides interoperability between existing state-of-the-art tools. As an open source Matlab toolbox, you can see in the small box that sparse reproducibility research tools represent the research community.