当前位置: X-MOL 学术Microprocess. Microsyst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Convolutional neural network-based restoration method of basketball contour image
Microprocessors and Microsystems ( IF 1.9 ) Pub Date : 2021-01-13 , DOI: 10.1016/j.micpro.2021.104004
Qiang Fu , Xiaoe Zhang , Huarui Li

Basketball image restoration is the process of taking damaged / noise images and predicting clean, original images. The vulnerability can take many forms such as motion blur, noise and camera misfocusing. Image Reconstruction Performed by this imaging point source, which is activated by converting blurred image, the so-called point diffusion function (including line) using the dot source image to recover the lost blurring process image information. The traditional outline tracking algorithm for basketball shooting dynamic hand image is vague, has poor stability and takes a long time. Recurrence nest tracking algorithm based on the dynamic boundary. The motion that the camera arm monitors are used to determine the target of the curve. The effective stiffness matrix is ​​obtained by initial calculation, as well as by using the characteristic curve recurrence calculation. The system image will then be applied to the dynamic boundary, where the energy is reduced to the target boundary. The purpose of basketball image restoration technology is to reduce noise and restore image processing technology's resolution loss in one of the image domain or frequency domains. Image restoration for basketball is performed on the frequency field except for the most direct previous art. It is computed by Fourier image and PSF, and the presence of convolution transforms the resolution loss caused by the blur factor. The probability sample is representing the entire population of sub-normal distribution with a Gaussian mixture model. The hybrid system, under normal conditions, which belongs to a subset of the data point seems obvious that this is a graded without learning is a subfield



中文翻译:

基于卷积神经网络的篮球轮廓图像恢复方法

篮球图像恢复是拍摄受损/噪点图像并预测干净原始图像的过程。该漏洞可能采取多种形式,例如运动模糊,噪点和相机对焦错误。图像重构由该成像点源执行,该成像点源通过转换模糊图像来激活,即使用点源图像恢复所谓的点扩散函数(包括线)以恢复丢失的模糊过程图像信息。传统的篮球动态手图像轮廓跟踪算法含糊不清,稳定性较差,需要较长时间。基于动态边界的递归嵌套跟踪算法。摄像机臂监视的运动用于确定曲线目标。有效刚度矩阵是通过初步计算得出的,以及通过使用特征曲线重复计算。然后,系统图像将应用于动态边界,在动态边界处,能量减少到目标边界。篮球图像恢复技术的目的是在图像域或频域之一中减少噪声并恢复图像处理技术的分辨率损失。除了最直接的现有技术外,篮球的图像恢复是在频域上执行的。它是由傅立叶图像和PSF计算的,卷积的存在改变了由模糊因子引起的分辨率损失。概率样本使用高斯混合模型表示次正态分布的总体。在正常条件下,混合动力系统

更新日期:2021-01-18
down
wechat
bug