Galaxies, human eyes, and artificial neural networks

The quantitative morphological classification of galaxies is important for understanding the origin of type frequency and correlations with environment. However, galaxy morphological classification is still mainly done visually by dedicated individuals, in the spirit of Hubble's original scheme and its modifications. The rapid increase in data on galaxy images at low and high redshift calls for a re-examination of the classification schemes and for automatic methods. Here are shown results from a systematic comparison of the dispersion among human experts classifying a uniformly selected sample of more than 800 digitized galaxy images. These galaxy images were then classified by six of the authors independently. The human classifications are compared with each other and with an automatic classification by an artificial neural network, which replicates the classification by a human expert to the same degree of agreement as that between two human experts.     

Leaf classification in sunflower crops by computer vision and neural networks

In this article, we present an automatic leaves image classification system for sunflower crops using neural networks, which could be used in selective herbicide applications. The system is comprised of four main stages. First, a segmentation based on rgb color space is performed. Second, many different features are detected and then extracted from the segmented image. Third, the most discriminable set of features are selected. Finally, the Generalized Softmax Perceptron (GSP) neural network architecture is used in conjunction with the recently proposed Posterior Probability Model Selection (PPMS) algorithm for complexity selection in order to select the leaves in an image and then classify them either as sunflower or non-sunflower. The experimental results show that the proposed system achieves a high level of accuracy with only five selected discriminative features obtaining an average Correct Classification Rate of 85% and an area under the receiver operation curve over 90%, for the test set.     

基于神经网络的公路运输量预测模型及应用

以福建省公路旅客周转量和货物周转量的统计资料为基础,结合神经网络技术原理,应用BP神经网络方法建立3维输入、单输出、隐层单元数为15的3层神经网络模型,分别对福建省公路旅客周转量和货物周转量进行预测.结果表明,各月的旅客周转量和货物量预测值的最大相对误差的绝对值分别为0.4890%和0.4495%.该模型具有简便实用、预测精度高的优点.     

利用双通道卷积神经网络的图像超分辨率算法

目的 图像超分辨率算法在实际应用中有着较为广泛的需求和研究。然而传统基于样本的超分辨率算法均使用简单的图像梯度特征表征低分辨率图像块,这些特征难以有效地区分不同的低分辨率图像块。针对此问题,在传统基于样本超分辨率算法的基础上,提出双通道卷积神经网络学习低分辨率与高分辨率图像块相似度进行图像超分辨率的算法。方法 首先利用深度卷积神经网络学习得到有效的低分辨率与高分辨率图像块之间相似性度量,然后根据输入低分辨率图像块与高分辨率图像块字典基元的相似度重构出对应的高分辨率图像块。结果 本文算法在Set5和Set14数据集上放大3倍情况下分别取得了平均峰值信噪比(PSNR)为32.53 dB与29.17 dB的效果。结论 本文算法从低分辨率与高分辨率图像块相似度学习角度解决图像超分辨率问题,可以更好地保持结果图像中的边缘信息,减弱结果中的振铃现象。本文算法可以很好地适用于自然场景图像的超分辨率增强任务。     

人工神经网络在水产科学中的应用

人工神经网络（ANN）是一种动态信息（处理）系统，它具有联想记忆、自组织、自适应、自学习和容错等优异的特性而得到广泛应用。ANN已广泛地应用于诸如模式识别、拟合、分类、决策和预测等领域，而水产科学有很多涉及上述技术的问题。本文在简述ANN结构和工作原理的基础上，讨论分析了利用BP神经网络模型、自组织特性神经网络或Kohonen神经网络模型进行分类、模式识别、图像处理和鉴别、预测与评价、系统模拟以及最优化和多目标决策等方面的应用实例。从神经网络模型建模和数据预处理原理研究了应用人工神经网络技术建模的局限性和缺陷。并明确指出：若不采用检验样本监控学习过程，对于一定数量的样本数据，过大的神经网络结构将不可避免地引起对样本数据的过拟合，从而得到了不能正确反映样本数据结构和内在特性和神经网络模型，而可能是对样本数据的噪声的反映。本文最后探讨了人工神经网络技术与模糊数学、逻辑控制和拓扑学以及非确定性原理相结合的应用趋势。     

Replicator neural networks for universal optimal source coding

Replicator neural networks self-organize by using their inputs as desired outputs; they internally form a compressed representation for the input data. A theorem shows that a class of replicator networks can, through the minimization of mean squared reconstruction error (for instance, by training on raw data examples), carry out optimal data compression for arbitrary data vector sources. Data manifolds, a new general model of data sources, are then introduced and a second theorem shows that, in a practically important limiting case, optimal-compression replicator networks operate by creating an essentially unique natural coordinate system for the manifold.     

人工神经网络在水产科学中的应用

人工神经网络（ANN）是一种动态信息（处理）系统，它具有联想记忆、自组织、自适应、自学习和容错等优异的特性而得到广泛应用。ANN已广泛地应用于诸如模式识别、拟合、分类、决策和预测等领域，而水产科学有很多涉及上述技术的问题。本文在简述ANN结构和工作原理的基础上，讨论分析了利用BP神经网络模型、自组织特性神经网络或Kohonen神经网络模型进行分类、模式识别、图像处理和鉴别、预测与评价、系统模拟以及最优化和多目标决策等方面的应用实例。从神经网络模型建模和数据预处理原理研究了应用人工神经网络技术建模的局限性和缺陷。并明确指出：若不采用检验样本监控学习过程，对于一定数量的样本数据，过大的神经网络结构将不可避免地引起对样本数据的过拟合，从而得到了不能正确反映样本数据结构和内在特性和神经网络模型，而可能是对样本数据的噪声的反映。本文最后探讨了人工神经网络技术与模糊数学、逻辑控制和拓扑学以及非确定性原理相结合的应用趋势。     

基于神经网络的土密实度检测

为了通过神经网络预测土密实度,搭建了土密实度检测装置,采集了土的激光图像,并提取了土的吸收系数、散射系数、激光图像的纹理特征参量和图像的灰度变化率共13个特征参数,通过SPSS降维处理,提取了5个主成分因子作为输入特征;并利用BP神经网络预测密实度.结果表明:经过31次测试后达到了误差要求,预测值与环刀法结果比较的平均绝对误差为7.14%和平均相对误差为7.71%,所建立的预测模型可行;最后通过试验对预测模型进行验证,其预测值与环刀法结果比较的平均绝对误差为8.62%和平均相对误差为8.76%,说明用神经网络预测土密实度是可行的.     

Reducing the dimensionality of data with neural networks

High-dimensional data can be converted to low-dimensional codes by training a multilayer neural network with a small central layer to reconstruct high-dimensional input vectors. Gradient descent can be used for fine-tuning the weights in such "autoencoder" networks, but this works well only if the initial weights are close to a good solution. We describe an effective way of initializing the weights that allows deep autoencoder networks to learn low-dimensional codes that work much better than principal components analysis as a tool to reduce the dimensionality of data.     

梯度神经网络的H-稳定性

给出总体极小化问题的神经网络模型 ,分析了网络的电路实现。证明了网络平稳点集合的全局 H-稳定性 ,从而有效地解决了网络的可靠运行问题     

基于L-M神经网络的储粮害虫分类识别

基于图像处理的储粮害虫检测过程中,需要解决多种害虫多特征、混合度大的综合分类问题.本文提出采用基于L-M算法的多层前馈神经网络对害虫进行分类识别.实验表明,该神经网络和害虫分类识别系统拟和程度很高,并且采用的L-M算法,在网络训练速度及识别精度方面,都优于传统的BP算法.因此基于L-M算法的神经网络在害虫的在线识别方面有应用价值.     

BP神经网络在农业植物分类识别中的应用

本文尝试利用BP算法对中国农业植物的56科进行分类识别的训练，其训练结果在精度范围内平均准确率达到90％，实现了科级上的植物分类的快速识别的目的，为植物分类识别工作的智能化提供了又一种思路。