First Exploration

• LIST OF ALL PDFs IN THE FIRST SAMPLE WITH A PATTERN OF BIG NUMBER AND SMALL NUMBER ON THE UP-RIGHT CORNER.

• All filenames are in the form of UCD_Lehmann_0000.jpg

• I represented words and numbers according to their position for the convinience of viewing if words are correcly catched by pdf parser engine. one example is put in ./notes

0208

Nothing words are not displayed horizontally.

0628

• target is at the bottom-right corner

$165.00 as 's' '165' '00' 's' and '00' are a little bit higher than '165' but they have the same size.

0629

Only $46.98 >> Only S A6 >>98

note: 0628 and 0629 have the very similar format, horizontal layout and only black and white color.

0946

$1.79 >> (nothing)
$19.95 >> 95

note: left bottom corner, and the price is twice larger than the surrounding words.

0985

$10.75 >> $ 15
$2.99 >> $999
1.64 >> 164.
8.75 >> 8/5
$1.16 >> $1 240
$3.59 >> $359
$38.75 >> $3875
$16.50 >> $1R50

note: those prices are listed in the individual boxes, it might be the reason why we have high probability to get them correctly.

1048

Only $28.69 >> nN y

note: the price is in a circle, 'Only $28.69' is larger than the surranding words.

1353

$32.95 >> nothing
$355.86 >> nothing

note: background color is very strong.

1721

31.95 >> 95
38.95 >> 95

note: those prices are on the top of the sticker-like block and much larger than other words.

1983

$1.79 >> $ 79
$19.95 >> $1 Q9>
$2.49 >> $ 499
$27.95 >> $ 95
$1.99 >> $ 90
$22.65 >> $ 22 65
$2.99 >> $ 1
$34.75 >> $3 475
$3.29 >> $ 290
$37.50 >> $3 7 50

note: prices are on the top left corner. This picture has only black and white color. It might be the reason why there are more correct numbers.

2504

$29.75 >> nothing
$35.95 >> =93
$25.75 >> nothing
$36.50 >> e>0

note: '$' is smaller than first two numbers. Those prices are surrounded by other words and larger than the surrounding words.

2531

$3.99 >> %©&
$46.75 >> & 7
$5.99 >> nothing
$68.50 >> % 5

note: target prices are surrounded by image patterns.

2535

$4.79 >> 8g 4 9

note: this file has only black and white color.

2601

$1.59 >> nothing
$17.95 >> $ 95
$3.79 >> nothing
$42.75 >> 3 75
$1.50 >> nothing
$16.20 >> q 20
$3.99 >> G9
$46.75 >> $4675
$4.99 >> 99
$56.50 >> $56˚0
$8.66 >> 86
$93.49 >> $9349
$6.24 >> 5624
$71.17 >> ST1|
$6.25 >> $65
$71.25 >> $7125
$8.20 >> $820
$93.48 >> $9348

note: there are 9 pairs of prices in the menu, the one on the top of each pair could not be obtained correcly.

2654

$5.99 >> $
$68.50 >> nothing

note: font style might influence the result of teserrect.

2692

$1.49 >> 49
$16.50 >> $16˚˚
$1.69 >> % 69
$18.50 >> $ 18 5 0
$1.59 >> nothing
$17.95 >> 1 7 95

note: when there is a single 1, it is always very thick.

2694

$1.79 >> nothing
$19.95 >> $7095
$1.99 >> &$0
$22.50 >> $ 50
$2.19 >> *2 19
$24.50 >> 5 50

note: it has very low possibility to catch 1's in the file.

2759

$10.95 >> @PRU
$30.75 >> $30 "5

3001

$2.29 >> £929

note: orange background

3387

$2.69 >> 3 ous 68 note: pink background

3875

note: tesserect only gets few words from top-right corner and leaves blank for everywhere else.

##NOTES: it might be better if we can show the actural size of the words in the plot.