Document Page Layout Analysis Document Page Layout Analysis - - PowerPoint PPT Presentation

▶
document page layout analysis document page layout
SMART_READER_LITE
LIVE PREVIEW

Document Page Layout Analysis Document Page Layout Analysis - - PowerPoint PPT Presentation

Oct 22, 2023 435 likes •1.23k views

Document Page Layout Analysis Document Page Layout Analysis Bhabatosh Chanda Electronics and Communication Sciences Unit Indian Statistical Institute Indian Statistical Institute Kolkata 700108, India Acknowledgement Acknowledgement Amit Das

slide-1
SLIDE 1

Document Page Layout Analysis Document Page Layout Analysis

Bhabatosh Chanda

Electronics and Communication Sciences Unit Indian Statistical Institute Indian Statistical Institute Kolkata 700108, India

slide-2
SLIDE 2

Acknowledgement Acknowledgement

  • Amit Das IIEST Sibpur

Amit Das, IIEST, Sibpur

  • Sekhar Mandal, IIEST, Sibpur

S j S h

  • Sanjoy Kumar Saha, Jadavpur Univeristy
  • Ranjan Mandal, Indian Statistical Institute

January 30, 2017 2 Indian Statistical Institute

slide-3
SLIDE 3

Outline Outline

  • Introduction
  • Projection method

– Zone content classification

  • Morphological operators

– Skew correction

  • Morphology based method

Morphology based method

  • Deep learning based method
  • Performance evaluation
  • Database: examples
  • Conclusion

January 30, 2017 3 Indian Statistical Institute

slide-4
SLIDE 4

Introduction Introduction

  • Problem description

Problem description

  • Motivation

I f f OCR

  • Improve performance of OCR
  • Data compression
  • Graphics recognition
  • Browsing and navigation
  • Physical and logical structure

January 30, 2017 4 Indian Statistical Institute

slide-5
SLIDE 5

Problem Description Problem Description

5 January 30, 2017 Indian Statistical Institute

slide-6
SLIDE 6

Objective Objective

6 January 30, 2017 Indian Statistical Institute

slide-7
SLIDE 7

Major Source of Document Pages Major Source of Document Pages

1 Books

  • 1. Books
  • 2. Journals

3 i

  • 3. Magazines
  • 4. Newspapers
  • 5. Forms and leaflets

6 Reports

  • 6. Reports

January 30, 2017 Indian Statistical Institute 7

slide-8
SLIDE 8

Types of document pages Types of document pages

Consider books and journals Consider books and journals

  • Title page

bli h ’

  • Publisher’s page
  • Table of Contents
  • Text page
  • Index page

Index page

January 30, 2017 Indian Statistical Institute 8

slide-9
SLIDE 9

Different types of pages Different types of pages

Title page Publisher’s page Title page Publisher s page

9 January 30, 2017 Indian Statistical Institute

slide-10
SLIDE 10

Different types of pages Different types of pages

Table of Content page Table of Content page Table of Content page Table of Content page

10 January 30, 2017 Indian Statistical Institute

slide-11
SLIDE 11

Different types of pages Different types of pages

Text page‐1 Text page‐2 Text page‐1 Text page‐2

11 January 30, 2017 Indian Statistical Institute

slide-12
SLIDE 12

Different types of pages Different types of pages

Text page‐3 Index page Text page‐3 Index page

12 January 30, 2017 Indian Statistical Institute

slide-13
SLIDE 13

Issues in document page scanning Issues in document page scanning

  • Resolution

Resolution

  • Back page impression

G l i

  • Granular noise
  • Blotted text (specially in old documents)
  • Bending of pages at the binding
  • Skew

Skew

(due to placement of the page in the scanner)

January 30, 2017 Indian Statistical Institute 13

slide-14
SLIDE 14

Entities of Document Page Entities of Document Page

  • Text

Text

– Body text

  • Line  Word  Character
  • Line  Word  Character

– Heading

  • Non text
  • Non‐text

– Half‐tone T bl – Table – Graphics or line drawing

January 30, 2017 Indian Statistical Institute 14

slide-15
SLIDE 15

Entities of Document Page Entities of Document Page

  • Each detected zone or block must be homogeneous

Each detected zone or block must be homogeneous in terms of content or entity

  • Each zone will be input to one of the suitable

p modules based on entity.

– OCR system – Image compressor – Vectorization system

  • Output of these modules may be compiled and

archived using suitable structure.

January 30, 2017 Indian Statistical Institute 15

slide-16
SLIDE 16

Geometrical / Physical structure Geometrical / Physical structure

Page Block Word c h Non‐ text Document Page Block Line Word . . . . . a r a c Block Line Word . . . . t e r s Page Line

16 January 30, 2017 Indian Statistical Institute

slide-17
SLIDE 17

Logical structure Logical structure

Document Text Non‐Text Normal High‐lighted lf i Normal High‐lighted Half‐tone (image) Line drawing Body Heading Sub‐ heading Abstract Graphics Table

17 January 30, 2017 Indian Statistical Institute

slide-18
SLIDE 18

Logical structure Logical structure

  • Different entities:

Different entities:

– Text (red box) – Halftone (green box) – Table (magenta box) – Line drawing (blue box)

  • Reading direction

(dark blue arrow)

  • Link between entities

(brown arrow)

18 January 30, 2017 Indian Statistical Institute

slide-19
SLIDE 19

Zone / block detection Zone / block detection

  • One of the simple way is Projection method.

One of the simple way is Projection method.

  • Algorithm

– Take horizontal (or vertical) projection of foreground Take horizontal (or vertical) projection of foreground

  • pixels. (may be implemented as pixel count)

– If there exists a characteristic change in projection profile, h i l ( i l) put a horizontal (resp. vertical) separator. – Take horizontal and vertical direction alternately. Continue until above condition is satisfied – Continue, until above condition is satisfied.

  • Works well for structured document, usually the pages of

technical journals, books, etc.

January 30, 2017 Indian Statistical Institute 19

slide-20
SLIDE 20

Projection Method: An Example Projection Method: An Example

20 January 30, 2017 Indian Statistical Institute

slide-21
SLIDE 21

Example (contd.) Example (contd.)

21 January 30, 2017 Indian Statistical Institute

slide-22
SLIDE 22

Example (contd.)

22 January 30, 2017 Indian Statistical Institute

slide-23
SLIDE 23

Example (contd.)

23 January 30, 2017 Indian Statistical Institute

slide-24
SLIDE 24

Problems of Projection method Problems of Projection method

  • Cannot say what each block contains until further

Cannot say what each block contains until further analysis.

– Extract features from a zone – Recognize the zone content using a classifier

  • Results are highly dependent even on small skew in

the scanned page.

January 30, 2017 Indian Statistical Institute 24

slide-25
SLIDE 25

Zone content recognition Zone content recognition

Features:

  • Black pixel ratio (no. of black pixel / zone area)
  • Horizontal transition (black to white) count
  • Vertical transition (black to white) count
  • Normalized mean length of horizontal black pixel run
  • Normalized mean length of vertical black pixel run
  • Normalized mean length of vertical black pixel run
  • Connected component ratio

Classifier:

  • Two‐class (text and non‐text)

SVM with RBF kernel (accuracy 94.89%)

January 30, 2017 Indian Statistical Institute 25

Duong, Emptoz, Côté: Features for Printed Document Image Analysis, ICPR 2002.

slide-26
SLIDE 26

Zone content recognition Zone content recognition

  • Functional classification of text blocks

u c o a c ass ca o o e b oc s

– Title / Heading, Sub‐heading, Body text …

  • Features:

– complexity (measured by entropy) – visibility values (or relative boldness) di i l (h i l d i l) – directional compactness (horizontal and vertical) – geometric characteristics (block height, width, etc.)

  • Classifier:

Classifier:

– K‐means clustering followed by min. distance classifier

Bres, Eglin, and Gafneux, Unsupervised Clustering of Text Entities in Heterogeneous Grey Level

January 30, 2017 Indian Statistical Institute 26

Documents, ICPR, 2002.

slide-27
SLIDE 27

Problems of Projection method Problems of Projection method

  • Cannot say what each block contains until further

Cannot say what each block contains until further analysis

– Extract features from a zone – Recognize the zone content using a classifier

  • Results are highly dependent even on small skew in

the scanned page

– Detecting base line of each text line of the document – Determining orientation (slope) angle of base line – Estimation overall skew of the document page

January 30, 2017 Indian Statistical Institute 27

slide-28
SLIDE 28

Processing Tool Processing Tool

  • Spatial domain operator that can handle

Spatial domain operator that can handle shape information directly

  • Mathematically well defined
  • Mathematically well defined
  • Neighborhood operator such that hardware

i l i h ld b i l implementation should be simple

January 30, 2017 Indian Statistical Institute 28

slide-29
SLIDE 29

Mathematical Morphology Mathematical Morphology

  • Mathematical morphological operators are

Mathematical morphological operators are good choice. Objects Objects

  • All characters, figures, drawing, i.e., black

components against white background components against white background Structuring element R l i fi

  • Regular geometric figures:

– mostly line segment, square, circle, etc.

January 30, 2017 Indian Statistical Institute 29

slide-30
SLIDE 30

Morphological Operations Morphological Operations

Set theoretic operations (including union, intersection, etc.):

  • 1. Dilation
  • 1. Dilation
  • 2. Erosion
  • 3. Opening
  • 4. Closing

30 January 30, 2017 Indian Statistical Institute

slide-31
SLIDE 31

Morphological operator: Dilation Morphological operator: Dilation

  • Expands the objects.

Orig.

p j

 

B b A a b a B A      , |

SE:

where A is an object and B is SE.

Circ‐5

  • Properties:

Commutative, associative

Circ‐9

associative, distributive (over union), increasing

Line‐ 19

g

31 January 30, 2017 Indian Statistical Institute

slide-32
SLIDE 32

Morphological operator: Erosion Morphological operator: Erosion

  • Shrinks the objects.

Orig.

j

 

A p B p B A     |

SE:

where A is an object and B is SE.

Circ‐5

  • Properties:

Distributive (over intersection), increasing

Circ‐9

increasing.

  • Dilation and erosion are dual.

Line‐ 19

32 January 30, 2017 Indian Statistical Institute

slide-33
SLIDE 33

Morphological operator: Opening Morphological operator: Opening

  • Removes objects or parts of it

Orig.

j p that cannot fit in SE.

SE:

 

B B A B A    

where A is an object and B is SE. P ti

Circ‐5

  • Properties:

Increasing, idempotent

Circ‐9

idempotent, anti-extensive.

  • It is a filter.

Line‐ 19

f

33 January 30, 2017 Indian Statistical Institute

slide-34
SLIDE 34

Morphological operator: Closing Morphological operator: Closing

  • Appends to objects parts of

Orig.

pp j p background if SE does not fit.

SE:

  B

B A B A    

where A is an object and B is SE. P ti

Circ‐5

  • Properties:

Increasing, idempotent, and extensive.

Circ‐9

  • It is a filter.
  • Opening & closing are dual.

Line‐ 19

34 January 30, 2017 Indian Statistical Institute

slide-35
SLIDE 35

Detecting base line Detecting base line

  • Close the original image

Orig.

Close the original image with line SE of suitable length.

SE:

  • Open the close image

with same line SE.

Close Line‐ 29

  • Detect black to white

transition in vertical

Cl‐Op Line‐ 29

scan.

B‐W trans trans.

35 January 30, 2017 Indian Statistical Institute

slide-36
SLIDE 36

Font Font

  • Traditionally in metal typesetting a font is a

Traditionally, in metal typesetting, a font is a particular size, weight and style of a typeface.

  • The weight of a particular font is the thickness of

The weight of a particular font is the thickness of the character outlines relative to their height.

  • Font size is measured in point unit
  • Font size is measured in point unit.

1 point in ...... is equal to ... typographic units 1/12 picas typographic units 1/12 picas imperial/US units 1/72 inch metric (SI) units 0 3528 mm metric (SI) units 0.3528 mm

January 30, 2017 Indian Statistical Institute 36

slide-37
SLIDE 37

Size related parameters Size related parameters

  • X‐height or corpus height

X height or corpus height

  • Ascender

d

  • Descender
  • Scan resolution (in dpi)

Scan resolution (in dpi)

  • Font style: bold, italics, ornamental

January 30, 2017 Indian Statistical Institute 37

slide-38
SLIDE 38

Skew correction: An example Skew correction: An example

38 January 30, 2017 Indian Statistical Institute

slide-39
SLIDE 39

Pages with complex layout Pages with complex layout

39 January 30, 2017 Indian Statistical Institute

slide-40
SLIDE 40

Morphological algorithm Morphological algorithm

  • Text region is composed of small objects (characters) placed in

g p j ( ) p regular interval.

  • Opening the image with small SE removes the thin object

t ( t k f h t ) b t h i i ifi t ff t parts (strokes of character), but has insignificant effect on large objects in half‐tone etc.

  • Closing the image with small SE fills in white holes in small

g g

  • bjects (space within and between character), but has

insignificant effect on large white space or half‐tone. Th diff b l d d d i hi hli h

  • Thus difference between closed and opened image highlights

the text region.

  • Difference image is thresholded to detect text region.

Difference image is thresholded to detect text region.

January 30, 2017 Indian Statistical Institute 40

slide-41
SLIDE 41

Morphological approach: An example Morphological approach: An example

(a) Original image (b) Closed image (c) Opened image QUESTION: Size of structuring element?

41 January 30, 2017 Indian Statistical Institute

slide-42
SLIDE 42

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

42 January 30, 2017 Indian Statistical Institute

slide-43
SLIDE 43

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

43 January 30, 2017 Indian Statistical Institute

slide-44
SLIDE 44

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

44 January 30, 2017 Indian Statistical Institute

slide-45
SLIDE 45

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

45 January 30, 2017 Indian Statistical Institute

slide-46
SLIDE 46

Deep learning Deep learning

  • Popular technique for unsupervised feature

Popular technique for unsupervised feature extraction for supervised applications – Ex. object recognition. j g

  • Utilizes HUGE number of instances to train relatively

simpler system to perform more complicated task. p y p p

  • Training samples may be outcome of controlled or

uncontrolled data acquisition.

  • Requires very high computational resources for

implementing a reasonably meaningful system.

January 30, 2017 Indian Statistical Institute 46 / 73

slide-47
SLIDE 47

Detect text area using CNN Detect text area using CNN

Input: A document image Output:Text / Non text area Input: A document image Output:Text / Non‐text area

47 January 30, 2017 Indian Statistical Institute

slide-48
SLIDE 48

Solution strategy Solution strategy

Transforming the problem into a classification Transforming the problem into a classification Problem.

  • Divide the Input image into MxM patches

Divide the Input image into MxM patches.

  • Input: Image patch of size MxM
  • Output: Text Non text and Ambiguous
  • Output: Text, Non text, and Ambiguous

– Text: if >80% of the patch has text – Non‐text: if <20% of the patch has text area

  • te t

0% o t e patc as te t a ea – Ambiguous: otherwise

January 30, 2017 Indian Statistical Institute 48

slide-49
SLIDE 49

Training data Training data

49/59 January 30, 2017 Indian Statistical Institute

slide-50
SLIDE 50

Prepare training data Prepare training data

INPUT: document images with manually labeled text area. g y

  • From each image, overlapping patches of size 100x100 are

taken (stride along x, y is 20) and resized to 50x50

  • From each image, overlapping patches of size 50x50 are taken

(stride along x, y is 10)

  • Each 50x50 patch is divided into 4 patches of size 25x25 and

Each 50x50 patch is divided into 4 patches of size 25x25 and are resized back to 50x50.

  • We get total number of 825670 patches of size 50x50 as

training data from 8 images. Label: as described before.

January 30, 2017 Indian Statistical Institute 50

slide-51
SLIDE 51

Training blocks: Example Training blocks: Example

January 30, 2017 Indian Statistical Institute 51

slide-52
SLIDE 52

Model description Model description

Input: 50x50 Patch of gray scale. Layer (type) Output Shape Param # Layer (type) Output Shape Param # ========================================================== Convolution2D(3x3 @8) (8, 48, 48) 80 MaxPooling2D(2x2) (8, 24, 24) Convolution2D(3x3 @6) (6, 22, 22) 438 Convolution2D(3x3 @4) (4, 20, 20) 220 Flatten (1600) Flatten (1600) Dense(7) (7) 11207 Activation(Sigmoid) (7) Dense(3) (3) 24 Dense(3) (3) 24 Activation(Softmax) (3) ========================================================== Total parameters: 11969 Total parameters: 11969

52/59 January 30, 2017 Indian Statistical Institute

slide-53
SLIDE 53

Model description Model description

53/59 January 30, 2017 Indian Statistical Institute

slide-54
SLIDE 54

Training the model Training the model

  • Number of epoch: 200

Number of epoch: 200

  • Batch size: 100

i 0 0

  • Learning Rate: 0.01
  • Learning weight decay: 0.95
  • Optimizer: Stochastic gradient descent
  • Loss function: Mean squared error

Loss function: Mean squared error

January 30, 2017 Indian Statistical Institute 54

slide-55
SLIDE 55

Testing Testing

  • Input: A test image

p g

  • Take 50x50 patch and submit it to the trained model
  • If predicted class is text, color that patch as pink.
  • If predicted class is non‐text, color the patch as white.
  • If predicted class is ambiguous, then

– Divide that patch into 4 patches of size 25x25 and resize to – Divide that patch into 4 patches of size 25x25 and resize to 50x50 and submitted to the model. – If that 50x50 patch is again ambiguous, then color that patch as yellow (Ideally it should be done recursively until patch as yellow (Ideally it should be done recursively until we get no ambiguous patch) – Else color the patch as according to text or non‐text class.

January 30, 2017 Indian Statistical Institute 55

slide-56
SLIDE 56

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

56 January 30, 2017 Indian Statistical Institute

slide-57
SLIDE 57

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

57 January 30, 2017 Indian Statistical Institute

slide-58
SLIDE 58

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

58 January 30, 2017 Indian Statistical Institute

slide-59
SLIDE 59

Results Results

Input test image Resultant (labeled) image Input test image Resultant (labeled) image

59 January 30, 2017 Indian Statistical Institute

slide-60
SLIDE 60

An improved network An improved network

32×32 25×25×96 5×5×96 4×4×256 2×2×256 [N T ] [Non‐Text] [Text] Convolution Convolution Classification Average pooling Average pooling

Wang, Wu, Coates and Ng, End‐to‐End Text Recognition with Convolutional Neural

January 30, 2017 Indian Statistical Institute 60

Wang, Wu, Coates and Ng, nd to nd Text Recognition with Convolutional Neural Networks, ICPR 2012.

slide-61
SLIDE 61

Comparative results Comparative results

Simpler system Wang et al Simpler system Wang et al.

January 30, 2017 Indian Statistical Institute 61

slide-62
SLIDE 62

Comparative results Comparative results

Simpler system Wang et al Simpler system Wang et al.

January 30, 2017 Indian Statistical Institute 62

slide-63
SLIDE 63

Benchmark database Benchmark database

  • UW‐I, II, III databeases, Developed at University of

UW I, II, III databeases, Developed at University of Washington, Seattle, USA in 1996.

  • Widely used earliest database with1620 pages

y p g

  • Zones contain text, non‐text such as halftone, line

drawing, math and chemical equation. g q

  • The database also contains

– Page condition file : skew angle, noise. – Page attribute file : dominant font and other content. – Page bounding box file : location and size of zones.

January 30, 2017 Indian Statistical Institute 63

http://isis‐data.science.uva.nl/events/dlia//datasets/uwash3.html

slide-64
SLIDE 64

Benchmark database Benchmark database

  • Mediateam document database

Mediateam document database

  • Developed at University of Oulu, Finland in1998.
  • One of the early databases containing

One of the early databases containing

Pattern type Samples Text 4811 Text 4811 Graphics 735 Image 161 Composite 219

January 30, 2017 Indian Statistical Institute 64

Duong, Emptoz, Côté: Features for Printed Document Image Analysis, ICPR 2002.

slide-65
SLIDE 65

Benchmark database Benchmark database

  • Pattern Recognition and Image Analysis (PRImA) Layout

a e ecog

  • a d

age a ys s ( ) ayou Analysis dataset

  • Developed at University of Salford, Manchester
  • 1240 ground‐truthed pages from magazines (1085 pages)

and technical journals (155 pages)

  • Used in following contests

– ICDAR 2015 Recognition of Documents with Complex Layouts (RDCL2015) (RDCL2015) – ICDAR2013 Historical Newspaper Layout Analysis (HNLA2013) – ICDAR2011 Historical Document Layout Analysis (HDLAC 2011)

January 30, 2017 Indian Statistical Institute 65

http://www.primaresearch.org/datasets/Layout_Analysis

slide-66
SLIDE 66

Benchmark database Benchmark database

  • Historical Newspaper dataset (ENP dataset)
  • Developed at University of Salford, Manchester in

Europeana Newspapers Project

  • 500 ground truthed pages covering
  • 500 ground‐truthed pages covering

– 13 languages (German, french, English, Estonian, etc.) – 17th, 18th, 19th and 20th centuries , ,

  • Contains (total regions 61,619) including

– 1,497 image zones – 208 table zones – 46,889 text zones

January 30, 2017 Indian Statistical Institute 66

Clausn Clausner er et. Al Al, The The ENP ENP Im Imag age and and Gr Ground

  • und Tr

Truth Da Datase set of

  • f hi

historic

  • rical

al new newspap paper, ICD ICDAR 2015. 2015.

slide-67
SLIDE 67

Performance evaluation Performance evaluation

  • A document page D may be represented as a m tuple.

p g y p p D = (E1, E2, …, Em) where Eis are entities such as text, tables, half‐tone, etc.

  • Each entity has a unique property denoted by Prop.(Ei).
  • Document page image domain X has n bounding boxes Bj

(j 1 n) with such that: (j=1,…, n) with such that:

for ) ( ) (

1 n j

k j B B ii X B i   

  ) .( Prop ) .( Prop such that

  • ne
  • nly

and

  • ne

exists there every For ) ( for ) (

i j k j

E B i j iii k j B B ii     

January 30, 2017 Indian Statistical Institute 67

) .( Prop ) .( Prop and background called is \ ) (

1 i n j

E B B X W iv  

slide-68
SLIDE 68

Performance evaluation Performance evaluation

68 January 30, 2017 Indian Statistical Institute

slide-69
SLIDE 69

Performance evaluation Performance evaluation

69 January 30, 2017 Indian Statistical Institute

slide-70
SLIDE 70

Performance evaluation Performance evaluation

70 January 30, 2017 Indian Statistical Institute

slide-71
SLIDE 71

Performance evaluation Performance evaluation

71 January 30, 2017 Indian Statistical Institute

slide-72
SLIDE 72

Performance evaluation Performance evaluation

  • Both model and object graphs are directed acyclic

graph.

  • Let us represent the model graph by

G = (V L ) GM = (VM, LM) where VM = {M0, M1, M2, . . . , Mn} represents the set

  • f nodes or vertices and LM represents set of links.

M

p

  • Note that Mj = (BBj, bbj,Ej) and Ljk = (Mj, Mk).
  • Similarly the object graph is represented by

G (V L ) Go = (Vo, Lo)

  • And Oj = (Bbj, Ej ) and Ljk = (Oj, Ok).
  • Finally graph matching algorithm is employed

Finally, graph matching algorithm is employed.

January 30, 2017 Indian Statistical Institute 72

slide-73
SLIDE 73

Performance evaluation Performance evaluation

Das, Saha and Chanda, An empirical measure of performance of document image as, Saha and Chanda, An empirical measure of performance of document image segmentation algorithm, IJDAR, Vol. 4(3), 2002.

73 January 30, 2017 Indian Statistical Institute

slide-74
SLIDE 74

Performance evaluation Performance evaluation

  • Relation between BBj and bbj in model (groundtruth):

j j

(g )

  • For good segmentation of object node:

k j bb BB bb BB

k j j j

  for and 

For good segmentation of object node:

i j j i j

O M BB Bb bb node matches node if  

  • The error measure:

(i) Correct classification (True positive) = #(Bbj ∩ BBi). (ii) F l l (F l iti ) #(Bb \ BB ) (ii) False alarm (False positive) = #(Bbj \ BBi). (iii) Mis‐classification (False negative) = #(bbi \ Bbj ).

January 30, 2017 Indian Statistical Institute 74

slide-75
SLIDE 75

Conclusion Conclusion

  • Presented a document image segmentation

Presented a document image segmentation method based on shape features

  • Used mathematical morphological operators
  • Used mathematical morphological operators
  • Necessary for OCR and data compression
  • System is useful for development of digital

library providing facilities for electronic storage, searching, navigation

January 30, 2017 Indian Statistical Institute 75

slide-76
SLIDE 76

References References

  • B. Chanda and D. Dutta Majumder, Digital Image Processing and Analysis,

Prentice Hall of India New Delhi 2000 Prentice Hall of India, New Delhi, 2000.

  • A. K. Das and B. Chanda, A fast algorithm for skew detection of document

images using morphology, Intl. J. Of Document Analysis and Recognition}, Vol.4, pp.109‐114, 2001.

  • A. K. Das, S. K. Saha and B. Chanda, An empirical measure of performance
  • f document image segmentation algorithm, Intl. J. on Document

Analysis and Recognition, Vol.4, pp.183‐190, 2002.

  • S Mandal A K Das and B Chanda A Simple and Effective Table
  • S. Mandal, A. K. Das and B. Chanda, A Simple and Effective Table

Detection System from Document Images, Int. J. on Document Analysis and Recognition, Vol.8, pp.172‐182, 2006.

January 30, 2017 Indian Statistical Institute 76

slide-77
SLIDE 77

Thank you

77 January 30, 2017 Indian Statistical Institute