A Brief Summary of Yann's "Gradient-Based Learning Applied to Document Recognition"
来源:互联网 发布:特朗普 金正恩 知乎 编辑:程序博客网 时间:2024/06/05 17:16
Paper Info:Gradient-Based Learning Applied to Document Recognition
YANN LECUN, MEMBER, IEEE, L´EONBOTTOU, YOSHUA BENGIO, AND PATRICK HAFFNER
I. Introduction
II. CNN for isolatedcharacter recognition
Features of Tradition Pattern Recognition:
1. hand-designedfeature extractor
2. trainable classifier
Problem: Images too large;topology of input (space or temporal correlations) ignored
Solution:
Using Convolutional Networks
Features: 1)local receptive fields 2)shared weight 3)spatial or temporalsubsampling(Once a feature has been detected, location less important)->LeNet-5
III. Results andcomparison with other methods
IV. Multimodule systems and graph transformer networks(GTN)
V. Multiple object recognition: HOS (The first method for character string recognition)
Isolated characters TO strings of characters
optimizing a global criterion
A now classical method for segmentation andrecognition—HOS
Good candidate locations for cuts can be found by locating minima in the vertical projection profile, or minima of the distance between the upper and lower contours of the word.
Structure of the Process
Question: What's the meaning of Interpretation graph?
Definitions in the paper:
The goal of the recognitiontransformer is to generate a graph, called the interpretation graph orrecognition graph that contains all the possible interpretations for all thepossible segmentations of the input.
The interpretation graph hasalmost the same structure as the segmentation graph, except that each arc isreplaced by a set of arcs from and to the same node.
VI. Global training for graph transformer networks
?global training? The whole process?
1.Viterbi training 2.discriminative Viterbitraining 3.Forward training 4.discriminative forward training 5.remarks
VII. Multiple object recognition: Space displacement neural network
No segmentation needed
Problems: Expensive; neighbors; notsize-normalized
Solution:Convolutional Networks- A replicatedconvolutional
network, also called an SDNN
A. Interpreting theOutput of an SDNN with a GTN
B. Experiments withSDNN
C. Global Training ofSDNN
D. Object Detection and Spotting with SDNN
VIII. Graph transformernetworks and transducers
IX. & X. Applications
(Online Handwriting recognition system and check reading system)
- A Brief Summary of Yann's "Gradient-Based Learning Applied to Document Recognition"
- 2016.4.16 Gradient-based learning applied to document recognition[待更]
- OpenCV编程实现LeCun论文(Gradient-Based Learning Applied to Document Recognition)中的CNN
- 论文提要“Gradient based Learning Applied to Document Reocognition”
- Introduction to Gradient-Based Learning
- Introduction to Gradient-Based Learning
- A Survey of Deep Learning Techniques Applied to Trading
- A Brief Summary of My Blog System
- A Brief Overview of Deep Learning
- A Brief Introduction of Deep Learning
- A Brief Summary of Towers of Hanoi Puzzle
- A Summary of Current Machine Learning Frameworks
- brief summary of local feature
- Random Ponderings - A Brief Overview of Deep Learning
- a brief history of maching learning(机器学习简史)
- A summary about gradient descent
- Eric S. Raymond五部曲之A Brief History of Hackerdom
- A brief overview of Nova’s new object model
- Atlas权重测试中的bug调试
- mount 命令详解
- 单点登录和权限系统结合注入到自己的系统中
- 集成电路中的assert和deassert应该如何翻译?
- 深度探索Linux操作系统第二章:系统构建和原理解析-工具编译
- A Brief Summary of Yann's "Gradient-Based Learning Applied to Document Recognition"
- 用 nodejs 做反向代理服务器
- The LLDB Debugger
- k-means聚类及C代码实现
- linux系统启动加载流程纪要
- Navicat Premium 11.0.16
- URAL 1905 Travel in Time BFS(广度搜索)
- hdu1728 逃离迷宫 bfs
- Android中asset文件夹与raw文件夹的区别深入解析