Image pattern recognition with neural network

Microsoft PowerPoint Le Dung Image Pattern Recognition with Neural Network 11 2011 pptx h ê đề dà h h lớ hChuyên đề dành cho lớp cao học C IMAGE PATTERN RECOGNITIONIMAGE PATTERN RECOGNITION WITH NEURAL NETWORKWITH NEURAL NETWORKWITH NEURAL NETWORKWITH NEURAL NETWORK Dr Lê Dũngg School of Electronics and Telecommunications Hanoi University of Science and TechnologyHanoi University of Science and Technology Hà nội 112011ộ TABLE OF CONTENTTABLE OF CONTENT huyên đề dành cho lớp cao họcC 2 TABLE OF.

Trang 1

IMAGE PATTERN RECOGNITION

WITH NEURAL NETWORK

Dr Lê Dũngg

School of Electronics and Telecommunications Hanoi University of Science and Technology

Hà nội 11/2011 ộ

Trang 2

TABLE OF CONTENT TABLE OF CONTENT

Part I: Design a real application using image Design a real application using image p pattern attern

recognition with recognition with n neural eural n network etwork

1 Automatic

1 Automatic Envelopes Classification System in the post Envelopes Classification System in the post office office

Part II: Image Pattern Recognition

1 Automatic

1 Automatic Envelopes Classification System in the post Envelopes Classification System in the post office office

2 Skin color detector with Neural Network

Part II: Image Pattern Recognition

+ Digital Image and Image Acquisition + Image Enhancement

+ Image Segmentation + Image Pattern Recognition

Part III: Recognition with Neural Network

+ Theory of Neural Network + Using Neural Network for Pattern Recognition

Trang 3

MỤC TIÊU CỦA CHUYÊN ĐỀ MỤC TIÊU CỦA CHUYÊN ĐỀ

Hiểu thế nào là “Mẫu ảnh” ( Image Pattern ) và nguyên

Hiểu thế nào là Mẫu ảnh ( Image Pattern ) và nguyên

lý nhận dạng mẫu ảnh ( Image Pattern Recognition ).

cho việc nhận dạng bằng mạng Nơron

cho việc nhận dạng bằng mạng Nơron.

nhận dạng mẫu ảnh.

dạng mẫu ảnh bằng mạng nơron.

Trang 4

TÀI LIÊU THAM KHẢO CHÍNH

“Digital Image Processing”

Barnd Jähne Spring Verlag 1995

“Neural Networks for Pattern Recognition”

Bishop, C.M.

Oxford University Press, 1995.

“Neural Network Design”

Martin T Hagan, Howard B Demuth, Mark Beale Thomson Learning, 1996.

“Gradient-based learning applied to document recognition”

LeCun, Y., Bottou, L., Bengio, Y., Haffner, P Proceedings of the IEEE vol 86 1998

Proceedings of the IEEE, vol 86, 1998.

Trang 5

AUTOMATIC ENVELOPES CLASSIFICATION SYSTEM

(Project : 010-2001-TCT-RDP-BC-26 at the Research Institute of Post and Telecommunication 2000-2001)

Collected envelopes

Control & Operation System Centre

Recognition result

Collected envelopes

control

Automatic Handwritten Postcode

(postcode)

Classification of

envelope types

control

My design

Recognition

Mechanical

design

Light CCD

camera

Mechanical System to do classification

Separate, Index &

envelopes

Put each of envelopes

on conveyer belt Conveyer belt

Trang 6

DESIGN THE PROCESS OF FUNCTIONS DESIGN THE PROCESS OF FUNCTIONS

xxxxx

Error (reject)

Segmentation (Neural network) Recognition Preprocessing

Image WxH

28117

xxxxx

Image WxH

Classification OK

Index i

CCD Camera

Light

Conveyer belt

Envelope

Multi-layers Feed-Forward Neural Network

Preprocessing Segmentation Recognition

- Grayscale convert

- Enhancement

- Segment the postcode area

- Segment 5 areas for

- Patterns normalization

- Recognition using

- Binary convert

- Subtraction

- Noise filter

- Find envelope frame

g each code number

- Segment 5 number patterns

a neural network

- Check result (OK or reject) p

- Rotate & Crop Postcode area

Trang 7

NEURAL NETWORK FOR PATTERN RECOGNITION

16x16

Input layer

Input layer : 256

Layer 2

Layer 2 : 106

Normalized

Pattern 16x16

Layer 3

7x7 5x5 16 16

3x3 3x3

Activation function

Layer 2 : 106 (Feature map)

Output layer

10

Log-Sigmoid

net

n

e

+

=

1

Layer 3 : 68 (Sub-sampling) Output layer : 10

0 1 2 3 4 5 6 7 8 9

Output layer Result

(classification) Total : 440 neurons

Trang 8

SOFTWARE MODULE (3) RECOGNIZING POSTCODE

SOFTWARE MODULE (3) : RECOGNIZING POSTCODE

Trang 9

SKIN COLOR DETECTOR

FOR A REAL-TIME HAND GESTURE RECOGNITION SYSTEM

2D Color Camera

Skin color detector

Skin color detection image

Hand feature extraction

Hand gesture recognition

Hand trackingg Human-Robot interaction Human-Computer interaction

Trang 10

THE OVERVIEW OF THE OVERVIEW OF MY MY COMBINING METHOD COMBINING METHOD THE OVERVIEW OF

THE OVERVIEW OF MY MY COMBINING METHOD COMBINING METHOD

Image frame

The

t ti ti l

PRNN

statistical skin color model

Training data set

+ Skin color patterns

Training Building

Collecting

+ Skin color patterns + Non-skin color patterns

Elliptical skin color

Step 1: Using the statistical skin color model

Æ Fast detection with not high accuracy (coarse work)

Æ Check all pi els of the image frame (coarse data)

The originality of the method: p model

Multi-layers

Feed-Æ Check all pixels of the image frame (coarse data)

Step 2: Using the pattern recognition neural network

Æ Detect with high accuracy (fine work)

Æ L t t th d t ti f th t ti ti l ki l d l

Forward Neural Network

Æ Learn to correct the detection errors of the statistical skin color model

Trang 11

DEFINE A PATTERN FOR SKIN COLOR DEFINE A PATTERN FOR SKIN COLOR PRNN* PRNN*

Skin detection result based on

Skin detection result based on the color information of a pixel

Non-skin

?

Pattern for PRNN with 9 pixels

Under examination pixel and its

8 neighborhood pixels

Skin

Non-skin Skin

8 neighborhood pixels

* PRNN P R i i N l N k

* PRNN – Pattern Recognition Neural Network

Trang 12

TỔNG KẾT CÁC KHÁI NIỆM TRONG PHẦN I

¾ Khái niệm mẫu ảnh trong nhận dạng.

¾ Khái niệm tạo và chuẩn hoá mẫu ảnh

¾ Khái niệm nhận dạng ệ ậ ạ g

¾ Khái niệm mạng nơron và huấn luyện mạng nơron

¾ Khái niệm thư viện mẫu ảnh để huấn luyện

¾ Khái niệm huấn luyện mạng nơron

Định dạng
Số trang	12
Dung lượng	916,48 KB