$im = imagecreatefrompng(IMAGES_DIRECTORY . $row['filename']);
}
}
$sql = "SELECT count(*) as c FROM ocrtrain\r\n\t\t\t\tWHERE fid = '{$fid}' and vid = '{$vid}' and bid = '{$bid}'";
$cc = $db->GetRow($sql);
if ($cc['c'] > 0) {
print T_("Found duplicate") . " {$fid} {$vid} {$bid}";
} else {
$row['width'] = imagesx($im);
$row['height'] = imagesy($im);
$image = crop($im, applytransforms($box, $row));
$a1 = kfill_modified($image, 5);
$a2 = remove_boundary_noise($a1, 2);
$timage = resize_bounding($a2);
$bimage = thinzs_np($timage);
$t = sector_distance($bimage);
$count++;
$sql = "INSERT INTO ocrtrain (ocrtid,val,f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,fid,vid,bid,kb)\r\n\t\t\t\t\tVALUES (NULL,'{$val}','{$t[0][1]}','{$t[0][2]}','{$t[0][3]}','{$t[0][4]}','{$t[0][5]}','{$t[0][6]}','{$t[0][7]}','{$t[0][8]}','{$t[0][9]}','{$t[0][10]}','{$t[0][11]}','{$t[0][12]}','{$t[1][1]}','{$t[1][2]}','{$t[1][3]}','{$t[1][4]}','{$fid}','{$vid}','{$bid}','{$kb}')";
$db->Execute($sql);
}
}
}
print T_("Trained") . ": {$count} " . T_("characters");
//generate kb
generate_kb($kb);
print T_("Generated KB");
}
if (isset($_GET['submit'])) {
//run process in background
$qid = intval($_GET['qid']);
$verifiers = " AND (";
/**
* Return what character based on the guess and the given kb
*
* @param image $image The image of the character
* @param int $btid The box type this came from
* @param int $qid The questionniare this box came from
*
* @return char The character detected
* @author Adam Zammit <*****@*****.**>
* @since 2010-10-18
*/
function ocr_guess($image, $btid, $qid)
{
include_once dirname(__FILE__) . '/../config.inc.php';
include_once dirname(__FILE__) . '/../db.inc.php';
global $db;
//remove speckles
$a1 = kfill_modified($image, 5);
//remove boundary lines
$a2 = remove_boundary_noise($a1, 2);
//resize the image to suit the OCR functions
$timage = resize_bounding($a2);
//thin the image to a skeleton
$bimage = thinzs_np($timage);
//extract the 16 features from the image
$f = sector_distance($bimage);
$sql = "SELECT val,\r\n\t\t\t(\r\n\t\t\t(pow(exp( - IFNULL((pow((m1-'{$f[0][1]}'),2)/(2*v1)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m2-'{$f[0][2]}'),2)/(2*v2)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m3-'{$f[0][3]}'),2)/(2*v3)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m4-'{$f[0][4]}'),2)/(2*v4)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m5-'{$f[0][5]}'),2)/(2*v5)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m6-'{$f[0][6]}'),2)/(2*v6)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m7-'{$f[0][7]}'),2)/(2*v7)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m8-'{$f[0][8]}'),2)/(2*v8)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m9-'{$f[0][9]}'),2)/(2*v9)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m10-'{$f[0][10]}'),2)/(2*v10)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m11-'{$f[0][11]}'),2)/(2*v11)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m12-'{$f[0][12]}'),2)/(2*v12)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m13-'{$f[1][1]}'),2)/(2*v13)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m14-'{$f[1][2]}'),2)/(2*v14)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m15-'{$f[1][3]}'),2)/(2*v15)),0)),2)) +\r\n\t\t\t(pow(exp( - IFNULL((pow((m16-'{$f[1][4]}'),2)/(2*v16)),0)),2))\r\n\t\t\t) as calc\r\n\t\tFROM ocrkbdata\r\n\t\tJOIN ocrkbboxgroup ON (ocrkbdata.kb = ocrkbboxgroup.kb AND ocrkbboxgroup.btid = '{$btid}' AND ocrkbboxgroup.qid = '{$qid}')\r\n\t\tORDER BY calc DESC";
$guess = $db->GetRow($sql);
//DEBUG
//print $sql . "<br/>";
return $guess['val'];
}
$im = imagecreatefromstring($row['image']);
} else {
$im = imagecreatefrompng(IMAGES_DIRECTORY . $row['filename']);
}
$row['width'] = imagesx($im);
$row['height'] = imagesy($im);
$box['tlx'] += BOX_EDGE;
$box['tly'] += BOX_EDGE;
$box['brx'] -= BOX_EDGE;
$box['bry'] -= BOX_EDGE;
$timage = crop($im, applytransforms($box, $row));
$ktimage = kfill_modified($timage, 5);
$ttimage = remove_boundary_noise($ktimage, 2);
$kttimage = resize_bounding($ttimage);
$i = thinzs_np($kttimage);
$t = sector_distance($i);
$sql = "INSERT INTO ocrtrain (ocrtid,val,f1,f2,f3,f4,f5,f6,f7,f8,f9,f10,f11,f12,f13,f14,f15,f16,fid,vid,bid,kb)\r\n VALUES (NULL,'{$val}','{$t[0][1]}','{$t[0][2]}','{$t[0][3]}','{$t[0][4]}','{$t[0][5]}','{$t[0][6]}','{$t[0][7]}','{$t[0][8]}','{$t[0][9]}','{$t[0][10]}','{$t[0][11]}','{$t[0][12]}','{$t[1][1]}','{$t[1][2]}','{$t[1][3]}','{$t[1][4]}','{$fid}','{$vid}','{$bid}','{$kb}')";
$db->Execute($sql);
$sql = "DELETE from ocrprocess\r\n\t\tWHERE ocrprocessid = '{$o['ocrprocessid']}'";
$db->Execute($sql);
print T_("Trained") . ": {$val} " . T_("to knowledge base") . ": {$kb}";
$db->CompleteTrans();
}
if ($completed == 1) {
$sql = "SELECT count(*) as c\r\n\t\tFROM ocrprocess\r\n\t\tWHERE kb = '{$kb}'";
$rs = $db->GetRow($sql);
if ($rs['c'] == 0) {
generate_kb($kb);
print T_("Generated KB");
} else {
print T_("Did not generate KB as not all records trained");