Detection and Segmentation of Detection and Segmentation of - - PowerPoint PPT Presentation

detection and segmentation of detection and segmentation
SMART_READER_LITE
LIVE PREVIEW

Detection and Segmentation of Detection and Segmentation of - - PowerPoint PPT Presentation

Detection and Segmentation of Detection and Segmentation of Touching Characters in Touching Characters in Mathematical Expressions Mathematical Expressions A. Nomura, A. Nomura, K. Michishita, K. Michishita, S. Uchida, S. Uchida, M.


slide-1
SLIDE 1

Detection and Segmentation of Detection and Segmentation of Touching Characters in Touching Characters in Mathematical Expressions Mathematical Expressions

  • A. Nomura,
  • A. Nomura,
  • K. Michishita,
  • K. Michishita,
  • S. Uchida,
  • S. Uchida,
  • M. Suzuki
  • M. Suzuki

Kyushu University, Japan Kyushu University, Japan

slide-2
SLIDE 2

Suzuki Lab. Kyushu Univ.

2

I ntroduction I ntroduction

slide-3
SLIDE 3

Suzuki Lab. Kyushu Univ.

3

OCR for mathematical document OCR for mathematical document

Aim

Recognition of ordinary texts

and mathematical expressions

Merits

Storage size reduction

bitmap image → ASCII codes

Search services

theorem search, definition search, …

Format conversion

from scanned image to LaTeX, XML, Mathematica Notebook, Braille, …

slide-4
SLIDE 4

Suzuki Lab. Kyushu Univ.

4

Large categories (> 500)

  • alphabets, numerals, Greek, operators,

parentheses, big symbols (e.g., “Σ”), …

Various fonts

  • roman, italic, calligraphic, …

Various sizes and positions

  • sub-/super-scripts, fractions, …

Touching characters

  • 50% and more misrecognitions were

due to touching characters

Hurdles Hurdles

slide-5
SLIDE 5

Suzuki Lab. Kyushu Univ.

5

Touching characters in math. expressions Touching characters in math. expressions

conventional segmentation techniques will fail

)

touching not only horizontally but also diagonally

slide-6
SLIDE 6

Suzuki Lab. Kyushu Univ.

6

Our purpose Our purpose

Development of a novel segmentation

technique for touching characters in mathematical expressions

And higher recognition accuracy

slide-7
SLIDE 7

Suzuki Lab. Kyushu Univ.

7

Outline of the Outline of the proposed proposed technique technique

slide-8
SLIDE 8

Suzuki Lab. Kyushu Univ.

8

Outline of the proposed technique (1) Outline of the proposed technique (1)

… …

all connected components in document (image data with initial recognition result)

clustering based on shape similarity

  • computational efficiency
  • neglect of trivial shape difference

centroids about 1/10 compression

slide-9
SLIDE 9

Suzuki Lab. Kyushu Univ.

9

single (separated) characters

Outline of the proposed technique (2) Outline of the proposed technique (2)

Detection of touching characters Segmentation of touching characters

… …

candidates of touching char.

centroids

slide-10
SLIDE 10

Suzuki Lab. Kyushu Univ.

10

Detail of detection procedure Detail of detection procedure

“Γ” extraction of two shape features

difference > threshold ?

feature values of standard “Γ”

connected component initial recognition result candidates of touching chars

yes

no

single characters

slide-11
SLIDE 11

Suzuki Lab. Kyushu Univ.

11

candidates of touching chars

Detail of segmentation procedure Detail of segmentation procedure

single characters match! match! residual

thickening

finding 1st char finding 2nd char

slide-12
SLIDE 12

Suzuki Lab. Kyushu Univ.

12

Notes Notes

1.

Single characters from the same document are utilized

  • Components of touching characters are usually

found as single characters in the document

  • Font-style/size adaptive separation is realized

2.

Recognition result is provided

3.

Tolerant to false positives

  • If no match is found, the candidate is rejected

as a single character

slide-13
SLIDE 13

Suzuki Lab. Kyushu Univ.

13

Experimental Experimental results results

slide-14
SLIDE 14

Suzuki Lab. Kyushu Univ.

14

Database Database

391 pages from 21 math. documents 140,000 characters in math. expressions

groundtruth was manually attached to

each character

characters in ordinary text parts were

excluded in our experiment

2,978 touching char images (~ 6,000 chars)

4.2% of all 140,000 characters

slide-15
SLIDE 15

Suzuki Lab. Kyushu Univ.

15

Detection result Detection result

touching 2864 single (false positive) 10002

detection result

False negatives were small (1.6%) False positives were large (but will be rejected in

the segmentation procedure)

touching 2978 single 134375 all characters in math. expression (about 140,000 characters) 50 false negative

slide-16
SLIDE 16

Suzuki Lab. Kyushu Univ.

16

Segmentation result Segmentation result

touching 2864 single (false positive) 10002 failure 1396 success (rejected as single) 9984 success 1468 detection result failure (forced separation) 18

50% of touching chars were successfully separated Forced separations were very small

segmentation result

slide-17
SLIDE 17

Suzuki Lab. Kyushu Univ.

17

Segmentation result Segmentation result

slide-18
SLIDE 18

Suzuki Lab. Kyushu Univ.

18

Effect on total recognition rate Effect on total recognition rate

Recognition rate of all 140,000 characters

92.9 % → 95.1 % (2.2% up)

The number of misrecognitions

9710 → 6792 (30% reduction)

the proposed technique is very meaningful !

slide-19
SLIDE 19

Suzuki Lab. Kyushu Univ.

19

Conclusion Conclusion

Detection and segmentation procedures for

touching characters in math. expressions were investigated

About 50% of touching characters were

successfully detected and separated

Total character recognition rate was

improved from 92.9 % to 95.1 %.