基于数据流计算图的卷积神经网络的理论和设计

近年来在许多信号处理应用领域中,深度卷积神经网络引起了学术界和工业界很大的关注,其中基于数据流图的深度学习网络Tensorflow框架得到了很多人的青睐。但在一些商业落地的研究和调查中发现,部分机构涉及一些自开发的计算单元,而它不被大型网络框架所支持,又出于技术的保密性往往需要自行修改工业界的深度学习框架,这就造成了以下情况, 其一是工业级大型代码框架极其复杂,各大库之间的调用很深且一般没有注释,不容易读懂和修改,其二是对某一个单一应用来说工业界绝大多数的代码都是有冗的,这就使得代码整体比较臃肿(厚),

pytorch应用之——纸币识别(一)

飞翔 2019-05-31 17:22
这里数据集一共有39620张,而且背景单一,所以纸币面值的识别不是一个很难的问题。我用resnet18(自己稍微改了一些结构,影响不大)去训练这个数据集,迭代24次可以达到99.96%的精度。

SE-Inception v3架构的模型搭建(keras代码实现)

Mark 2018-09-28 16:47
2017年ImagNet冠军架构得主的精髓之处SENet架构(Squeeze And Excitation),关于细节处不再多说,只是该架构的基本结构图,和代码实现。并且代码实现此处是googleNet的Inception v3架构为基础加上SE的结构。

人民币面值识别

Flower_In_The_Dream 2019-08-04 22:11
人民币面值识别

人民币面值识别(附源码)-InceptionResNetV2

李劳(Felix) 2019-06-02 15:26
人民币面值识别, InceptionResNetV2

TinyMind人民币编码识别挑战赛-经验分享

mind_programmonkey 2019-05-31 20:54
简单卷积神经网络实现人民币实现

Very Deep Convolutional Networks for Large-Scale Image Recognition

辉仔 2018-01-10 10:31
In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of in

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

辉仔 2018-01-10 10:49
Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the trai

Rethinking the Inception Architecture for Computer Vision

辉仔 2018-01-10 11:10
Convolutional networks are at the core of most state-of-the-art computer vision solutions for a wide variety of tasks. Since 2014 very deep convolutional networks started to become mainstream, yieldin

A Neural Algorithm of Artistic Style

dwSun 2018-02-07 17:07
In fine art, especially painting, humans have mastered the skill to create unique visual experiences through composing a complex interplay between the content and style of an image. Thus far the algor
关注微信公众号