Does lstm_choice_mode only work for lstm?
35 views
Skip to first unread message
un C
Aug 12, 2020, 5:33:54 AM8/12/20
to tesseract-ocr
I am trying to get the second choice from Tesseract and I find the "lstm_choice_mode" parameter.
But it appears that it only work for lstm engine.(Only lstm gives more than one output for each characters)
Can anyone give me a clue?
Results for lstm engine:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<title></title>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8"/>
<meta name='ocr-system' content='tesseract v5.0.0-alpha.20200328' />
<meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word ocrp_wconf'/>
</head>
<body>
<div class='ocr_page' id='page_1' title='image "C:\Users\YUNTSE~1\AppData\Local\Temp\tess_4yg1i978.PNG"; bbox 0 0 224 114; ppageno 0'>
<div class='ocr_carea' id='block_1_1' title="bbox 35 37 154 66">
<p class='ocr_par' id='par_1_1' lang='chi_tra' title="bbox 35 37 154 66">
<span class='ocr_line' id='line_1_1' title="bbox 35 37 154 66; baseline 0.017 -2; x_size 36; x_descenders 9; x_ascenders 9">
<span class='ocrx_word' id='word_1_1' title='bbox 35 39 91 65; x_wconf 92'>固定
<span class='ocrx_cinfo' id='lstm_choices_1_1_1'>
<span class='ocrx_cinfo' id='choice_1_1_1' title='x_confs 0'>固</span></span>
<span class='ocrx_cinfo' id='lstm_choices_1_1_2'>
<span class='ocrx_cinfo' id='choice_1_1_2' title='x_confs 95.486488'>定</span>
<span class='ocrx_cinfo' id='choice_1_1_3' title='x_confs 0'>守</span>
<span class='ocrx_cinfo' id='choice_1_1_4' title='x_confs 0'>放</span>
<span class='ocrx_cinfo' id='choice_1_1_5' title='x_confs 0'>軍</span>
<span class='ocrx_cinfo' id='choice_1_1_6' title='x_confs 0'>說</span>
<span class='ocrx_cinfo' id='choice_1_1_7' title='x_confs 0'>補</span></span>
</span>
<span class='ocrx_word' id='word_1_2' title='bbox 109 37 154 66; x_wconf 96'>資產
<span class='ocrx_cinfo' id='lstm_choices_1_2_1'>
<span class='ocrx_cinfo' id='choice_1_2_1' title='x_confs 0'>資</span>
<span class='ocrx_cinfo' id='choice_1_2_2' title='x_confs 0'>謂</span>
<span class='ocrx_cinfo' id='choice_1_2_3' title='x_confs 0'>寬</span>
<span class='ocrx_cinfo' id='choice_1_2_4' title='x_confs 0'>寅</span>
<span class='ocrx_cinfo' id='choice_1_2_5' title='x_confs 0'>宮</span></span>
<span class='ocrx_cinfo' id='lstm_choices_1_2_2'>
<span class='ocrx_cinfo' id='choice_1_2_6' title='x_confs 96.327087'>產</span>
<span class='ocrx_cinfo' id='choice_1_2_7' title='x_confs 0'>廬</span>
<span class='ocrx_cinfo' id='choice_1_2_8' title='x_confs 0'>宙</span>
<span class='ocrx_cinfo' id='choice_1_2_9' title='x_confs 0'>寬</span>
<span class='ocrx_cinfo' id='choice_1_2_10' title='x_confs 0'>放</span></span>
</span>
</span>
</p>
</div>
</div>
</body>
</html>Results for legacy engine:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<title></title>
<meta http-equiv="Content-Type" content="text/html;charset=utf-8"/>
<meta name='ocr-system' content='tesseract v5.0.0-alpha.20200328' />
<meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word ocrp_wconf'/>
</head>
<body>
<div class='ocr_page' id='page_1' title='image "C:\Users\YUNTSE~1\AppData\Local\Temp\tess_s7_zf3u2.PNG"; bbox 0 0 224 114; ppageno 0'>
<div class='ocr_carea' id='block_1_1' title="bbox 35 37 154 66">
<p class='ocr_par' id='par_1_1' lang='business_tax_payment' title="bbox 35 37 154 66">
<span class='ocr_line' id='line_1_1' title="bbox 35 37 154 66; baseline 0 -0.999; x_size 36; x_descenders 9; x_ascenders 9">
<span class='ocrx_word' id='word_1_1' title='bbox 35 37 154 66; x_wconf 38'>固定固定產
</span>
</span>
</p>
</div>
</div>
</body>
</html>Reply all
Reply to author
Forward
0 new messages