An efficient controller structure with minimum roundoff noise gain

Based on a polynomial operator approach, a new sparse controller structure is derived, which is actually an improved version of the recently proposed structure [Li, G. (2004). A polynomial-operator-based DFIIt structure for IIR filters. IEEE Transactions on Circuits and Systems II, 51, 147-151]. The performance of the proposed structure is analyzed by deriving the corresponding expression of roundoff noise gain and the problem of finding optimized sparse structures is solved efficiently with a genetic algorithm (GA). A numerical example is given, which shows that the newly developed structure can achieve much better performance than some well-known structures and particularly outperforms the traditional optimal fully parametrized realization greatly in terms of reducing roundoff noise and implementation complexity.     

Fisher discrimination based low rank matrix recovery for face recognition

In this paper, we consider the issue of computing low rank (LR) recovery of matrices with sparse errors. Based on the success of low rank matrix recovery in statistical learning, computer vision and signal processing, a novel low rank matrix recovery algorithm with Fisher discrimination regularization (FDLR) is proposed. Standard low rank matrix recovery algorithm decomposes the original matrix into a set of representative basis with a corresponding sparse error for modeling the raw data. Motivated by the Fisher criterion, the proposed FDLR executes low rank matrix recovery in a supervised manner, i.e., taking the with-class scatter and between-class scatter into account when the whole label information are available. The paper shows that the formulated model can be solved by the augmented Lagrange multipliers and provides additional discriminating power over the standard low rank recovery models. The representative bases learned by the proposed method are encouraged to be closer within the same class, and as far as possible between different classes. Meanwhile, the sparse error recovered by FDLR is not discarded as usual, but treated as a feedback in the following classification tasks. Numerical simulations demonstrate that the proposed algorithm achieves the state of the art results.     

基于线性动态系统稀疏编码的异常事件检测

线性动态系统模型结合稀疏编码实现异常事件检测。线性动态系统可有效地捕捉动态纹理在时间和空间的转移信息,描述视频的时空小块。然而,线性动态系统属于非欧氏空间,无法直接用传统的稀疏编码进行异常检测。基于约束凸优化公式,将相似性变换与稀疏编码结合,可实现线性动态系统稀疏编码的优化求解。实验表明,所提出的方法具有更好的性能。     

Video summarization via minimum sparse reconstruction

The rapid growth of video data demands both effective and efficient video summarization methods so that users are empowered to quickly browse and comprehend a large amount of video content. In this paper, we formulate the video summarization task with a novel minimum sparse reconstruction (MSR) problem. That is, the original video sequence can be best reconstructed with as few selected keyframes as possible. Different from the recently proposed convex relaxation based sparse dictionary selection method, our proposed method utilizes the true sparse constraint L₀ norm, instead of the relaxed constraint _L2,1$L_{2,1}$ norm, such that keyframes are directly selected as a sparse dictionary that can well reconstruct all the video frames. An on-line version is further developed owing to the real-time efficiency of the proposed MSR principle. In addition, a percentage of reconstruction (POR) criterion is proposed to intuitively guide users in obtaining a summary with an appropriate length. Experimental results on two benchmark datasets with various types of videos demonstrate that the proposed methods outperform the state of the art.     

低功耗异构计算架构的高光谱遥感图像分类研究

高光谱图像分类算法通常需要逐点对图像中的像素点进行迭代处理,计算复杂度及并行程度存在较大差异。随着高光谱遥感图像空间、光谱和辐射分辨率的不断提升,这些算法无法满足实时处理海量遥感图像数据的需求。通过分析NPU存储计算一体化模式与遥感图像分类算法的实现步骤,设计低功耗CPU+NPU异构资源计算架构的低秩稀疏子空间聚类（LRSSC）算法,将数据密集型计算转移至NPU,并利用NPU数据驱动并行计算和内置AI加速,对基于机器学习算法的海量遥感数据进行实时分类。受到big.LITTLE计算范式的启发,CPU+NPU异构资源计算架构由8 bit和低精度位宽NPU共同组成以提高整体吞吐量,同时减少图网络推理过程中的能量损耗。实验结果表明,与CPU计算架构和CPU+GPU异构计算架构的LRSSC算法相比,CPU+NPU异构计算架构的LRSSC算法在Pavia University遥感数据集下的计算速度提升了3~14倍。     

Text detection in images using sparse representation with discriminative dictionaries

Text detection is important in the retrieval of texts from digital pictures, video databases and webpages. However, it can be very challenging since the text is often embedded in a complex background. In this paper, we propose a classification-based algorithm for text detection using a sparse representation with discriminative dictionaries. First, the edges are detected by the wavelet transform and scanned into patches by a sliding window. Then, candidate text areas are obtained by applying a simple classification procedure using two learned discriminative dictionaries. Finally, the adaptive run-length smoothing algorithm and projection profile analysis are used to further refine the candidate text areas. The proposed method is evaluated on the Microsoft common test set, the ICDAR 2003 text locating set, and an image set collected from the web. Extensive experiments show that the proposed method can effectively detect texts of various sizes, fonts and colors from images and videos.     

Sparse polynomial division using a heap

In 1974, Johnson showed how to multiply and divide sparse polynomials using a binary heap. This paper introduces a new algorithm that uses a heap to divide with the same complexity as multiplication. It is a fraction-free method that also reduces the number of integer operations for divisions of polynomials with integer coefficients over the rationals. Heap-based algorithms use very little memory and do not generate garbage. They can run in the CPU cache and achieve high performance. We compare our C implementation of sparse polynomial multiplication and division with integer coefficients to the routines of the Magma, Maple, Pari, Singular and Trip computer algebra systems.