Tesseract parameters in 3.02 version

parameters in 3.02 version

name	value	description
editor_image_xpos	590	Editor image X Pos
editor_image_ypos	10	Editor image Y Pos
editor_image_menuheight	50	Add to image height for menu bar
editor_image_word_bb_color	7	Word bounding box colour
editor_image_blob_bb_color	4	Blob bounding box colour
editor_image_text_color	2	Correct text colour
editor_dbwin_xpos	50	Editor debug window X Pos
editor_dbwin_ypos	500	Editor debug window Y Pos
editor_dbwin_height	24	Editor debug window height
editor_dbwin_width	80	Editor debug window width
editor_word_xpos	60	Word window X Pos
editor_word_ypos	510	Word window Y Pos
editor_word_height	240	Word window height
editor_word_width	655	Word window width
textord_debug_tabfind	0	Debug tab finding
textord_debug_bugs	0	Turn on output related to bugs in tab finding
textord_testregion_left	-1	Left edge of debug reporting rectangle
textord_testregion_top	-1	Top edge of debug reporting rectangle
textord_testregion_right	2147483647	Right edge of debug rectangle
textord_testregion_bottom	2147483647	Bottom edge of debug rectangle
textord_tabfind_show_partitions	0	Show partition bounds, waiting if >1
devanagari_split_debuglevel	0	Debug level for split shiro-rekha process.
edges_max_children_per_outline	10	Max number of children inside a character outline
edges_max_children_layers	5	Max layers of nested children inside a character outline
edges_children_per_grandchild	10	Importance ratio for chucking outlines
edges_children_count_limit	45	Max holes allowed in blob
edges_min_nonhole	12	Min pixels for potential char in box
edges_patharea_ratio	40	Max lensq/area for acceptable child outline
edges_maxedgelength	16000	Max steps in any outline
textord_fp_chop_error	2	Max allowed bending of chop cells
textord_tabfind_show_images	0	Show image blobs
textord_skewsmooth_offset	2	For smooth factor
textord_skewsmooth_offset2	1	For smooth factor
textord_test_x	-1	coord of test pt
textord_test_y	-1	coord of test pt
textord_min_blobs_in_row	4	Min blobs before gradient counted
textord_spline_minblobs	8	Min blobs in each spline segment
textord_spline_medianwin	6	Size of window for spline segmentation
textord_max_blob_overlaps	4	Max number of blobs a big blob can overlap
textord_min_xheight	10	Min credible pixel xheight
textord_lms_line_trials	12	Number of linew fits to do
oldbl_holed_losscount	10	Max lost before fallback line used
pitsync_linear_version	6	Use new fast algorithm
pitsync_fake_depth	1	Max advance fake generation
textord_tabfind_show_strokewidths	0	Show stroke widths
textord_dotmatrix_gap	3	Max pixel gap for broken pixed pitch
textord_debug_block	0	Block to do debug on
textord_pitch_range	2	Max range test on pitch
textord_words_veto_power	5	Rows required to outvote a veto
wordrec_display_segmentations	0	Display Segmentations
classify_radius_gyr_min_man	255	Minimum Radius of Gyration Mantissa 0-255:
classify_radius_gyr_min_exp	0	Minimum Radius of Gyration Exponent 0-255:
classify_radius_gyr_max_man	158	Maximum Radius of Gyration Mantissa 0-255:
classify_radius_gyr_max_exp	8	Maximum Radius of Gyration Exponent 0-255:
classify_num_cp_levels	3	Number of Class Pruner Levels
image_default_resolution	300	Image resolution dpi
equationdetect_save_bi_image	0	Save input bi image
equationdetect_save_spt_image	0	Save special character image
equationdetect_save_seed_image	0	Save the seed image
equationdetect_save_merged_image	0	Save the merged image
textord_debug_images	0	Use greyed image background for debug
textord_debug_printable	0	Make debug windows printable
textord_space_size_is_variable	0	If true, word delimiter spaces are assumed to have variable width, even though characters have fixed pitch.
textord_tabfind_show_initial_partitions	0	Show partition bounds
textord_tabfind_show_reject_blobs	0	Show blobs rejected as noise
textord_tabfind_show_columns	0	Show column bounds
textord_tabfind_show_blocks	0	Show final block bounds
textord_tabfind_find_tables	1	run table detection
textord_tabfind_show_color_fit	0	Show stroke widths
devanagari_split_debugimage	0	Whether to create a debug image for split shiro-rekha process.
textord_show_fixed_cuts	0	Draw fixed pitch cell boundaries
edges_use_new_outline_complexity	0	Use the new outline complexity module
edges_debug	0	turn on debugging for this module
edges_children_fix	0	Remove boxy parents of char-like children
gapmap_debug	0	Say which blocks have tables
gapmap_use_ends	0	Use large space at start and end of rows
gapmap_no_isolated_quanta	0	Ensure gaps not less than 2quanta wide
textord_heavy_nr	0	Vigorously remove noise
textord_show_initial_rows	0	Display row accumulation
textord_show_parallel_rows	0	Display page correlated rows
textord_show_expanded_rows	0	Display rows after expanding
textord_show_final_rows	0	Display rows after final fitting
textord_show_final_blobs	0	Display blob bounds after pre-ass
textord_test_landscape	0	Tests refer to land/port
textord_parallel_baselines	1	Force parallel baselines
textord_straight_baselines	0	Force straight baselines
textord_old_baselines	1	Use old baseline algorithm
textord_old_xheight	0	Use old xheight algorithm
textord_fix_xheight_bug	1	Use spline baseline
textord_fix_makerow_bug	1	Prevent multiple baselines
textord_debug_xheights	0	Test xheight algorithms
textord_biased_skewcalc	1	Bias skew estimates with line length
textord_interpolating_skew	1	Interpolate across gaps
textord_new_initial_xheight	1	Use test xheight mechanism
textord_really_old_xheight	0	Use original wiseowl xheight
textord_oldbl_debug	0	Debug old baseline generation
textord_debug_baselines	0	Debug baseline generation
textord_oldbl_paradef	1	Use para default mechanism
textord_oldbl_split_splines	1	Split stepped splines
textord_oldbl_merge_parts	1	Merge suspect partitions
oldbl_corrfix	1	Improve correlation of heights
oldbl_xhfix	0	Fix bug in modes threshold for xheights
textord_ocropus_mode	0	Make baselines for ocropus
textord_tabfind_only_strokewidths	0	Only run stroke widths
textord_tabfind_vertical_text	1	Enable vertical detection
textord_tabfind_force_vertical_text	0	Force using vertical text page mode
textord_tabfind_vertical_horizontal_mix	1	find horizontal lines such as headers in vertical page mode
textord_tabfind_show_initialtabs	0	Show tab candidates
textord_tabfind_show_finaltabs	0	Show tab vectors
textord_dump_table_images	0	Paint table detection output
textord_show_tables	0	Show table regions
textord_tablefind_show_mark	0	Debug table marking steps in detail
textord_tablefind_show_stats	0	Show page stats used in table finding
textord_tablefind_recognize_tables	0	Enables the table recognizer for table layout and filtering.
textord_all_prop	0	All doc is proportial text
textord_debug_pitch_test	0	Debug on fixed pitch test
textord_disable_pitch_test	0	Turn off dp fixed pitch algorithm
textord_fast_pitch_test	0	Do even faster pitch algorithm
textord_debug_pitch_metric	0	Write full metric stuff
textord_show_row_cuts	0	Draw row-level cuts
textord_show_page_cuts	0	Draw page-level cuts
textord_pitch_cheat	0	Use correct answer for fixed/prop
textord_blockndoc_fixed	0	Attempt whole doc/block fixed pitch
textord_show_initial_words	0	Display separate words
textord_show_new_words	0	Display separate words
textord_show_fixed_words	0	Display forced fixed pitch words
textord_blocksall_fixed	0	Moan about prop blocks
textord_blocksall_prop	0	Moan about fixed pitch blocks
textord_blocksall_testing	0	Dump stats when moaning
textord_test_mode	0	Do current test
textord_pitch_scalebigwords	0	Scale scores on big words
textord_restore_underlines	1	Chop underlines & put back
textord_fp_chopping	1	Do fixed pitch chopping
textord_force_make_prop_words	0	Force proportional word segmentation on all rows
textord_chopper_test	0	Chopper is being tested.
wordrec_display_all_blobs	0	Display Blobs
wordrec_display_all_words	0	Display Words
wordrec_blob_pause	0	Blob pause
poly_debug	0	Debug old poly
poly_wide_objects_better	1	More accurate approx on wide things
wordrec_display_splits	0	Display splits
editor_image_win_name	EditorImage	Editor image window name
editor_dbwin_name	EditorDBWin	Editor debug window name
editor_word_name	BlnWords	BL normalized word window
editor_debug_config_file		Config file to apply to single words
fx_debugfile	FXDebug	Name of debugfile
classify_font_name	UnknownFont	Default font name to be used in training
classify_training_file	MicroFeatures	Training file
debug_file		File to send tprintf output to
textord_underline_threshold	0.5	Fraction of width occupied
edges_childarea	0.5	Min area fraction of child outline
edges_boxarea	0.875	Min area fraction of grandchild for box
textord_fp_chop_snap	0.5	Max distance of chop pt from vertex
gapmap_big_gaps	1.75	xht multiplier
textord_spline_shift_fraction	0.02	Fraction of line spacing for quad
textord_spline_outlier_fraction	0.1	Fraction of line spacing for outlier
textord_skew_ile	0.5	Ile of gradients for page skew
textord_skew_lag	0.01	Lag for skew on row accumulation
textord_linespace_iqrlimit	0.2	Max iqr/median for linespace
textord_width_limit	8	Max width of blobs to make rows
textord_chop_width	1.5	Max width before chopping
textord_expansion_factor	1	Factor to expand rows by in expand_rows
textord_overlap_x	0.5	Fraction of linespace for good overlap
textord_minxh	0.25	fraction of linesize for min xheight
textord_min_linesize	1.25	* blob height for initial linesize
textord_excess_blobsize	1.3	New row made if blob makes row this big
textord_occupancy_threshold	0.4	Fraction of neighbourhood
textord_underline_width	2	Multiple of line_size for underline
textord_min_blob_height_fraction	0.75	Min blob height/top to include blob top into xheight stats
textord_xheight_mode_fraction	0.4	Min pile height to make xheight
textord_ascheight_mode_fraction	0.08	Min pile height to make ascheight
textord_descheight_mode_fraction	0.08	Min pile height to make descheight
textord_ascx_ratio_min	1.25	Min cap/xheight
textord_ascx_ratio_max	1.8	Max cap/xheight
textord_descx_ratio_min	0.25	Min desc/xheight
textord_descx_ratio_max	0.6	Max desc/xheight
textord_xheight_error_margin	0.1	Accepted variation
oldbl_xhfract	0.4	Fraction of est allowed in calc
oldbl_dot_error_size	1.26	Max aspect ratio of a dot
textord_oldbl_jumplimit	0.15	X fraction for new partition
pitsync_joined_edge	0.75	Dist inside big blob for chopping
pitsync_offset_freecut_fraction	0.25	Fraction of cut for free cuts
textord_tabfind_vertical_text_ratio	0.5	Fraction of textlines deemed vertical to use vertical page mode
textord_tabfind_aligned_gap_fraction	0.75	Fraction of height used as a minimum gap for aligned blobs.
textord_tabvector_vertical_gap_fraction	0.5	max fraction of mean blob width allowed for vertical gaps in vertical text
textord_tabvector_vertical_box_ratio	0.5	Fraction of box matches required to declare a line vertical
textord_projection_scale	0.2	Ding rate for mid-cuts
textord_balance_factor	1	Ding rate for unbalanced char cells
textord_wordstats_smooth_factor	0.05	Smoothing gap stats
textord_width_smooth_factor	0.1	Smoothing width stats
textord_words_width_ile	0.4	Ile of blob widths for space est
textord_words_maxspace	4	Multiple of xheight
textord_words_default_maxspace	3.5	Max believable third space
textord_words_default_minspace	0.6	Fraction of xheight
textord_words_min_minspace	0.3	Fraction of xheight
textord_words_default_nonspace	0.2	Fraction of xheight
textord_words_initial_lower	0.25	Max inital cluster size
textord_words_initial_upper	0.15	Min initial cluster spacing
textord_words_minlarge	0.75	Fraction of valid gaps needed
textord_words_pitchsd_threshold	0.04	Pitch sync threshold
textord_words_def_fixed	0.016	Threshold for definite fixed
textord_words_def_prop	0.09	Threshold for definite prop
textord_pitch_rowsimilarity	0.08	Fraction of xheight for sameness
words_initial_lower	0.5	Max inital cluster size
words_initial_upper	0.15	Min initial cluster spacing
words_default_prop_nonspace	0.25	Fraction of xheight
words_default_fixed_space	0.75	Fraction of xheight
words_default_fixed_limit	0.6	Allowed size variance
textord_words_definite_spread	0.3	Non-fuzzy spacing region
textord_spacesize_ratiofp	2.8	Min ratio space/nonspace
textord_spacesize_ratioprop	2	Min ratio space/nonspace
textord_fpiqr_ratio	1.5	Pitch IQR/Gap IQRthreshold
textord_max_pitch_iqr	0.2	Xh fraction noise in pitch
textord_fp_min_width	0.5	Min width of decent blobs
textord_underline_offset	0.1	Fraction of x to ignore
classify_cp_angle_pad_loose	45	Class Pruner Angle Pad Loose
classify_cp_angle_pad_medium	20	Class Pruner Angle Pad Medium
classify_cp_angle_pad_tight	10	CLass Pruner Angle Pad Tight
classify_cp_end_pad_loose	0.5	Class Pruner End Pad Loose
classify_cp_end_pad_medium	0.5	Class Pruner End Pad Medium
classify_cp_end_pad_tight	0.5	Class Pruner End Pad Tight
classify_cp_side_pad_loose	2.5	Class Pruner Side Pad Loose
classify_cp_side_pad_medium	1.2	Class Pruner Side Pad Medium
classify_cp_side_pad_tight	0.6	Class Pruner Side Pad Tight
classify_pp_angle_pad	45	Proto Pruner Angle Pad
classify_pp_end_pad	0.5	Proto Prune End Pad
classify_pp_side_pad	2.5	Proto Pruner Side Pad
classify_min_slope	0.414214	Slope below which lines are called horizontal
classify_max_slope	2.41421	Slope above which lines are called vertical
classify_norm_adj_midpoint	32	Norm adjust midpoint …
classify_norm_adj_curl	2	Norm adjust curl …
classify_pico_feature_length	0.05	Pico Feature Length
speckle_large_max_size	0.3	Max large speckle size
speckle_small_penalty	10	Small speckle penalty
speckle_large_penalty	10	Large speckle penalty
speckle_small_certainty	-1	Small speckle certainty
ambigs_debug_level	0	Debug level for unichar ambiguities
tessedit_single_match	0	Top choice only from CP
classify_debug_level	0	Classify debug level
classify_norm_method	1	Normalization Method …
matcher_debug_level	0	Matcher Debug Level
matcher_debug_flags	0	Matcher Debug Flags
classify_learning_debug_level	0	Learning Debug Level:
matcher_permanent_classes_min	1	Min # of permanent classes
matcher_min_examples_for_prototyping	3	Reliable Config Threshold
matcher_sufficient_examples_for_prototyping	5	Enable adaption even if the ambiguities have not been seen
classify_adapt_proto_threshold	230	Threshold for good protos during adaptive 0-255
classify_adapt_feature_threshold	230	Threshold for good features during adaptive 0-255
classify_class_pruner_threshold	229	Class Pruner Threshold 0-255
classify_class_pruner_multiplier	30	Class Pruner Multiplier 0-255:
classify_cp_cutoff_strength	7	Class Pruner CutoffStrength:
classify_integer_matcher_multiplier	14	Integer Matcher Multiplier 0-255:
il1_adaption_test	0	Dont adapt to i/I at beginning of word
dawg_debug_level	0	Set to 1 for general debug info, to 2 for more details, to 3 to see all the debug messages
hyphen_debug_level	0	Debug level for hyphenated words.
max_viterbi_list_size	10	Maximum size of viterbi list.
stopper_smallword_size	2	Size of dict word to be treated as non-dict word
stopper_debug_level	0	Stopper debug level
tessedit_truncate_wordchoice_log	10	Max words to keep in list
fragments_debug	0	Debug character fragments
segment_debug	0	Debug the whole segmentation process
max_permuter_attempts	10000	Maximum number of different character choices to consider during permutation. This limit is especially useful when user patterns are specified, since overly generic patterns can result in dawg search exploring an overly large number of options.
wordrec_num_seg_states	30	Segmentation states
repair_unchopped_blobs	1	Fix blobs that aren’t chopped
chop_debug	0	Chop debug
chop_split_length	10000	Split Length
chop_same_distance	2	Same distance
chop_min_outline_points	6	Min Number of Points on Outline
chop_inside_angle	-50	Min Inside Angle Bend
chop_min_outline_area	2000	Min Outline Area
chop_x_y_weight	3	X / Y length weight
segment_adjust_debug	0	Segmentation adjustment debug
wordrec_debug_level	0	Debug level for wordrec
segsearch_debug_level	0	SegSearch debug level
segsearch_max_pain_points	2000	Maximum number of pain points stored in the queue
segsearch_max_futile_classifications	10	Maximum number of pain point classifications per word thatdid not result in finding a better word choice.
language_model_debug_level	0	Language model debug level
language_model_ngram_order	8	Maximum order of the character ngram model
language_model_viterbi_list_max_num_prunable	10	Maximum number of prunable (those for which PrunablePath() is true) entries in each viterbi list recorded in BLOB_CHOICEs
language_model_viterbi_list_max_size	500	Maximum size of viterbi lists recorded inBLOB_CHOICEs
language_model_min_compound_length	3	Minimum length of compound words
language_model_fixed_length_choices_depth	3	Depth of blob choice lists to explore when fixed length dawgs are on
tessedit_pageseg_mode	6	Page seg mode: 0=osd only, 1=auto+osd, 2=auto, 3=col, 4=block, 5=line, 6=word, 7=char (Values from PageSegMode enum in publictypes.h)
tessedit_ocr_engine_mode	0	Which OCR engine(s) to run (Tesseract, Cube, both). Defaults to loading and running only Tesseract (no Cube,no combiner). Values from OcrEngineMode enum in tesseractclass.h)
pageseg_devanagari_split_strategy	0	Whether to use the top-line splitting process for Devanagari documents while performing page-segmentation.
ocr_devanagari_split_strategy	0	Whether to use the top-line splitting process for Devanagari documents while performing ocr.
bidi_debug	0	Debug level for BiDi
applybox_debug	1	Debug level
applybox_page	0	Page number to apply boxes from
tessedit_bigram_debug	0	Amount of debug output for bigram correction.
debug_x_ht_level	0	Reestimate debug
quality_min_initial_alphas_reqd	2	alphas in a good word
tessedit_tess_adaption_mode	39	Adaptation decision algorithm for tess
tessedit_test_adaption_mode	3	Adaptation decision algorithm for tess
paragraph_debug_level	0	Print paragraph debug info.
cube_debug_level	0	Print cube debug info.
tessedit_preserve_min_wd_len	2	Only preserve wds longer than this
crunch_rating_max	10	For adj length in rating per ch
crunch_pot_indicators	1	How many potential indicators needed
crunch_leave_lc_strings	4	Dont crunch words with long lower case strings
crunch_leave_uc_strings	4	Dont crunch words with long lower case strings
crunch_long_repetitions	3	Crunch words with long repetitions
crunch_debug	0	As it says
fixsp_non_noise_limit	1	How many non-noise blbs either side?
fixsp_done_mode	1	What constitues done for spacing
debug_fix_space_level	0	Contextual fixspace debug
x_ht_acceptance_tolerance	8	Max allowed deviation of blob top outside of font data
x_ht_min_change	8	Min change in xht before actually trying it
suspect_level	99	Suspect marker level
suspect_space_level	100	Min suspect level for rejecting spaces
suspect_short_words	2	Dont Suspect dict wds longer than this
tessedit_reject_mode	0	Rejection algorithm
tessedit_ok_mode	5	Acceptance decision algorithm
tessedit_image_border	2	Rej blbs near image edge limit
min_sane_x_ht_pixels	8	Reject any x-ht lt or eq than this
tessedit_page_number	-1	-1 -> All pages , else specifc page to process
tessdata_manager_debug_level	0	Debug level for TessdataManager functions.
tosp_debug_level	0	Debug data
tosp_enough_space_samples_for_median	3	or should we use mean
tosp_redo_kern_limit	10	No.samples reqd to reestimate for row
tosp_few_samples	40	No.gaps reqd with 1 large gap to treat as a table
tosp_short_row	20	No.gaps reqd with few cert spaces to use certs
tosp_sanity_method	1	How to avoid being silly
textord_max_noise_size	7	Pixel size of noise
textord_noise_sizefraction	10	Fraction of size for maxima
textord_noise_translimit	16	Transitions for normal blob
textord_noise_sncount	1	super norm blobs to save row
use_definite_ambigs_for_classifier	0	Use definite ambiguities when running character classifier
use_ambigs_for_adaption	0	Use ambigs for deciding whether to adapt to a character
prioritize_division	0	Prioritize blob division over chopping
classify_enable_learning	1	Enable adaptive classifier
tess_cn_matching	0	Character Normalized Matching
tess_bn_matching	0	Baseline Normalized Matching
classify_enable_adaptive_matcher	1	Enable adaptive classifier
classify_use_pre_adapted_templates	0	Use pre-adapted classifier templates
classify_save_adapted_templates	0	Save adapted templates to a file
classify_enable_adaptive_debugger	0	Enable match debugger
disable_character_fragments	1	Do not include character fragments in the results of the classifier
classify_debug_character_fragments	0	Bring up graphical debugging windows for fragments training
matcher_debug_separate_windows	0	Use two different windows for debugging the matching: One for the protos and one for the features.
classify_bln_numeric_mode	0	Assume the input is numbers [0-9].
load_system_dawg	1	Load system word dawg.
load_freq_dawg	1	Load frequent word dawg.
load_unambig_dawg	1	Load unambiguous word dawg.
load_punc_dawg	1	Load dawg with punctuation patterns.
load_number_dawg	1	Load dawg with number patterns.
load_fixed_length_dawgs	1	Load fixed length dawgs (e.g. for non-space delimited languages)
load_bigram_dawg	1	Load dawg with special word bigrams.
use_only_first_uft8_step	0	Use only the first UTF8 step of the given string when computing log probabilities.
stopper_no_acceptable_choices	0	Make AcceptableChoice() always return false. Useful when there is a need to explore all segmentations
save_raw_choices	1	Save all explored raw choices
permute_debug	0	Debug char permutation process
permute_script_word	0	Turn on word script consistency permuter
segment_segcost_rating	0	incorporate segmentation cost in word rating?
segment_nonalphabetic_script	0	Don’t use any alphabetic-specific tricks.Set to true in the traineddata config file for scripts that are cursive or inherently fixed-pitch
permute_fixed_length_dawg	0	Turn on fixed-length phrasebook search permuter
permute_chartype_word	0	Turn on character type (property) consistency permuter
save_doc_words	0	Save Document Words
doc_dict_enable	1	Enable Document Dictionary
ngram_permuter_activated	0	Activate character-level n-gram-based permuter
permute_only_top	0	Run only the top choice permuter
merge_fragments_in_matrix	1	Merge the fragments in the ratings matrix and delete them after merging
wordrec_no_block	0	Don’t output block information
wordrec_enable_assoc	1	Associator Enable
force_word_assoc	0	force associator to run regardless of what enable_assoc is.This is used for CJK where component grouping is necessary.
fragments_guide_chopper	0	Use information from fragments to guide chopping process
chop_enable	1	Chop enable
chop_vertical_creep	0	Vertical creep
assume_fixed_pitch_char_segment	0	include fixed-pitch heuristics in char segmentation
use_new_state_cost	0	use new state cost heuristics for segmentation state evaluation
wordrec_debug_blamer	0	Print blamer debug messages
wordrec_run_blamer	0	Try to set the blame for errors
enable_new_segsearch	0	Enable new segmentation search path.
save_alt_choices	1	Save alternative paths found during chopping and segmentation search
language_model_ngram_on	0	Turn on/off the use of character ngram model
language_model_ngram_use_only_first_uft8_step	0	Use only the first UTF8 step of the given string when computing log probabilities.
language_model_ngram_space_delimited_language	1	Words are delimited by space
language_model_use_sigmoidal_certainty	0	Use sigmoidal score for certainty
tessedit_resegment_from_boxes	0	Take segmentation and labeling from box file
tessedit_resegment_from_line_boxes	0	Conversion of word/line box file to char box file
tessedit_train_from_boxes	0	Generate training data from boxed chars
tessedit_make_boxes_from_boxes	0	Generate more boxes from boxed chars
tessedit_dump_pageseg_images	0	Dump intermediate images made during page segmentation
tessedit_ambigs_training	0	Perform training for ambiguities
tessedit_adapt_to_char_fragments	1	Adapt to words that contain a character composed form fragments
tessedit_adaption_debug	0	Generate and print debug information for adaption
applybox_learn_chars_and_char_frags_mode	0	Learn both character fragments (as is done in the special low exposure mode) as well as unfragmented characters.
applybox_learn_ngrams_mode	0	Each bounding box is assumed to contain ngrams. Only learn the ngrams whose outlines overlap horizontally.
tessedit_display_outwords	0	Draw output words
tessedit_training_tess	0	Call Tess to learn blobs
tessedit_dump_choices	0	Dump char choices
tessedit_fix_fuzzy_spaces	1	Try to improve fuzzy spaces
tessedit_unrej_any_wd	0	Dont bother with word plausibility
tessedit_fix_hyphens	1	Crunch double hyphens?
tessedit_redo_xheight	1	Check/Correct x-height
tessedit_enable_doc_dict	1	Add words to the document dictionary
tessedit_debug_fonts	0	Output font info per char
tessedit_debug_block_rejection	0	Block and Row stats
tessedit_enable_bigram_correction	1	Enable correction based on the word bigram dictionary.
debug_acceptable_wds	0	Dump word pass/fail chk
tessedit_tess_adapt_to_rejmap	0	Use reject map to control Tesseract adaption
tessedit_minimal_rej_pass1	0	Do minimal rejection on pass 1 output
tessedit_test_adaption	0	Test adaption criteria
tessedit_matcher_log	0	Log matcher activity
save_blob_choices	0	Save the results of the recognition step (blob_choices) within the correspondingWERD_CHOICE
test_pt	0	Test for point
docqual_excuse_outline_errs	0	Allow outline errs in unrejection?
tessedit_good_quality_unrej	1	Reduce rejection on good docs
tessedit_use_reject_spaces	1	Reject spaces?
tessedit_preserve_blk_rej_perfect_wds	1	Only rej partially rejected words in block rejection
tessedit_preserve_row_rej_perfect_wds	1	Only rej partially rejected words in row rejection
tessedit_dont_blkrej_good_wds	0	Use word segmentation quality metric
tessedit_dont_rowrej_good_wds	0	Use word segmentation quality metric
tessedit_row_rej_good_docs	1	Apply row rejection to good docs
tessedit_reject_bad_qual_wds	1	Reject all bad quality wds
tessedit_debug_doc_rejection	0	Page stats
tessedit_debug_quality_metrics	0	Output data to debug file
bland_unrej	0	unrej potential with no chekcs
unlv_tilde_crunching	1	Mark v.bad words for tilde crunch
crunch_early_merge_tess_fails	1	Before word crunch?
crunch_early_convert_bad_unlv_chs	0	Take out ~^ early?
crunch_terrible_garbage	1	As it says
crunch_pot_garbage	1	POTENTIAL crunch garbage
crunch_leave_ok_strings	1	Dont touch sensible strings
crunch_accept_ok	1	Use acceptability in okstring
crunch_leave_accept_strings	0	Dont pot crunch sensible strings
crunch_include_numerals	0	Fiddle alpha figures
tessedit_prefer_joined_punct	0	Reward punctation joins
tessedit_write_block_separators	0	Write block separators in output
tessedit_write_rep_codes	0	Write repetition char code
tessedit_write_unlv	0	Write .unlv output file
tessedit_create_hocr	0	Write .html hOCR output file
suspect_constrain_1Il	0	UNLV keep 1Il chars rejected
tessedit_minimal_rejection	0	Only reject tess failures
tessedit_zero_rejection	0	Dont reject ANYTHING
tessedit_word_for_word	0	Make output have exactly one word perWERD
tessedit_zero_kelvin_rejection	0	Dont reject ANYTHING ATALL
tessedit_consistent_reps	1	Force all rep chars the same
tessedit_rejection_debug	0	Adaption debug
tessedit_flip_0O	1	Contextual 0O O0 flips
rej_trust_doc_dawg	0	Use DOC dawg in 11l conf. detector
rej_1Il_use_dict_word	0	Use dictword test
rej_1Il_trust_permuter_type	1	Dont double check
rej_use_tess_accepted	1	Individual rejection control
rej_use_tess_blanks	1	Individual rejection control
rej_use_good_perm	1	Individual rejection control
rej_use_sensible_wd	0	Extend permuter check
rej_alphas_in_number_perm	0	Extend permuter check
tessedit_create_boxfile	0	Output text with boxes
tessedit_write_images	0	Capture the image from the IPE
interactive_display_mode	0	Run interactively?
tessedit_override_permuter	1	According to dict_word
textord_tabfind_show_vlines	0	Debug line finding
textord_use_cjk_fp_model	0	Use CJK fixed pitch model
tessedit_init_config_only	0	Only initialize with the config file. Useful if the instance is not going to be used forOCR but say only for layout analysis.
textord_equation_detect	0	Turn on equation detector
textord_single_height_mode	0	Script has no xheight, so use a single mode
tosp_old_to_method	0	Space stats use prechopping?
tosp_old_to_constrain_sp_kn	0	Constrain relative values of inter and intra-word gaps for old_to_method.
tosp_only_use_prop_rows	1	Block stats to use fixed pitch rows?
tosp_force_wordbreak_on_punct	0	Force word breaks on punct to break long lines in non-space delimited langs
tosp_use_pre_chopping	0	Space stats use prechopping?
tosp_old_to_bug_fix	0	Fix suspected bug in old code
tosp_block_use_cert_spaces	1	Only stat OBVIOUSspaces
tosp_row_use_cert_spaces	1	Only stat OBVIOUSspaces
tosp_narrow_blobs_not_cert	1	Only stat OBVIOUSspaces
tosp_row_use_cert_spaces1	1	Only stat OBVIOUSspaces
tosp_recovery_isolated_row_stats	1	Use row alone when inadequate cert spaces
tosp_only_small_gaps_for_kern	0	Better guess
tosp_all_flips_fuzzy	0	Pass ANY flip to context?
tosp_fuzzy_limit_all	1	Dont restrict kn->sp fuzzy limit to tables
tosp_stats_use_xht_gaps	1	Use within xht gap for wd breaks
tosp_use_xht_gaps	1	Use within xht gap for wd breaks
tosp_only_use_xht_gaps	0	Only use within xht gap for wd breaks
tosp_rule_9_test_punct	0	Dont chng kn to space next to punct
tosp_flip_fuzz_kn_to_sp	1	Default flip
tosp_flip_fuzz_sp_to_kn	1	Default flip
tosp_improve_thresh	0	Enable improvement heuristic
textord_no_rejects	0	Don’t remove noise blobs
textord_show_blobs	0	Display unsorted blobs
textord_show_boxes	0	Display unsorted blobs
textord_noise_rejwords	1	Reject noise-like words
textord_noise_rejrows	1	Reject noise-like rows
textord_noise_debug	0	Debug row garbage detector
m_data_sub_dir	tessdata/	Directory for data files
classify_learn_debug_str		Class str to debug learning
user_words_suffix		A list of user-provided words.
user_patterns_suffix		A list of user-provided patterns.
output_ambig_words_file		Output file for ambiguities found in the dictionary
word_to_debug		Word for which stopper debug information should be printed to stdout
word_to_debug_lengths		Lengths of unichars in word_to_debug
tessedit_char_blacklist		Blacklist of chars not to recognize
tessedit_char_whitelist		Whitelist of chars to recognize
tessedit_write_params_to_file		Write all parameters to the given file.
applybox_exposure_pattern	.exp	Exposure value follows this pattern in the image filename. The name of the image files are expected to be in the form [lang].[fontname].exp[num].tif
chs_leading_punct	(’`”	Leading punctuation
chs_trailing_punct1	).,;:?!	1st Trailing punctuation
chs_trailing_punct2	)’`”	2nd Trailing punctuation
outlines_odd	%|	Non standard number of outlines
outlines_2	ij!?%”:;	Non standard number of outlines
numeric_punctuation	.,	Punct. chs expectedWITHIN numbers
unrecognised_char	|	Output char for unidentified blobs
ok_repeated_ch_non_alphanum_wds	-?*=	Allow NN to unrej
conflict_set_I_l_1	Il1[]	Il1 conflict set
file_type	.tif	Filename extension
tessedit_load_sublangs		List of languages to load with this one
classify_char_norm_range	0.2	Character Normalization Range …
classify_min_norm_scale_x	0	Min char x-norm scale …
classify_max_norm_scale_x	0.325	Max char x-norm scale …
classify_min_norm_scale_y	0	Min char y-norm scale …
classify_max_norm_scale_y	0.325	Max char y-norm scale …
matcher_good_threshold	0.125	Good Match (0-1)
matcher_great_threshold	0	Great Match (0-1)
matcher_perfect_threshold	0.02	Perfect Match (0-1)
matcher_bad_match_pad	0.15	Bad Match Pad (0-1)
matcher_rating_margin	0.1	New template margin (0-1)
matcher_avg_noise_size	12	Avg. noise blob length
matcher_clustering_max_angle_delta	0.015	Maximum angle delta for prototype clustering
classify_misfit_junk_penalty	0	Penalty to apply when a non-alnum is vertically out of its expected textline position
rating_scale	1.5	Rating scaling factor
certainty_scale	20	Certainty scaling factor
tessedit_class_miss_scale	0.00390625	Scale factor for features not used
classify_character_fragments_garbage_certainty_threshold	-3	Exclude fragments that do not look like whole characters from training and adaption
segment_penalty_dict_frequent_word	1	Score multiplier for word matches which have good case andare frequent in the given language (lower is better).
segment_penalty_dict_case_ok	1.1	Score multiplier for word matches that have good case (lower is better).
segment_penalty_dict_case_bad	1.3125	Default score multiplier for word matches, which may have case issues (lower is better).
segment_penalty_ngram_best_choice	1.24	Multipler to for the best choice from the ngram model.
segment_penalty_dict_nonword	1.25	Score multiplier for glyph fragment segmentations which do not match a dictionary word (lower is better).
segment_penalty_garbage	1.5	Score multiplier for poorly cased strings that are not in the dictionary and generally look like garbage (lower is better).
certainty_scale	20	Certainty scaling factor
stopper_nondict_certainty_base	-2.5	Certainty threshold for non-dict words
stopper_phase2_certainty_rejection_offset	1	Reject certainty offset
stopper_certainty_per_char	-0.5	Certainty to add for each dict char above small word size.
stopper_allowable_character_badness	3	Max certaintly variation allowed in a word (in sigma)
stopper_ambiguity_threshold_gain	8	Gain factor for ambiguity threshold.
stopper_ambiguity_threshold_offset	1.5	Certainty offset for ambiguity threshold.
bestrate_pruning_factor	2	Multiplying factor of current best rate to prune other hypotheses
segment_reward_script	0.95	Score multipler for script consistency within a word. Being a ‘reward’ factor, it should be <= 1. Smaller value implies bigger reward.
segment_reward_chartype	0.97	Score multipler for char type consistency within a word.
segment_reward_ngram_best_choice	0.99	Score multipler for ngram permuter’s best choice (only used in the Han script path).
doc_dict_pending_threshold	0	Worst certainty for using pending dictionary
doc_dict_certainty_threshold	-2.25	Worst certainty for words that can be inserted into thedocument dictionary
wordrec_worst_state	1	Worst segmentation state
tessedit_certainty_threshold	-2.25	Good blob limit
chop_split_dist_knob	0.5	Split length adjustment
chop_overlap_knob	0.9	Split overlap adjustment
chop_center_knob	0.15	Split center adjustment
chop_sharpness_knob	0.06	Split sharpness adjustment
chop_width_change_knob	5	Width change adjustment
chop_ok_split	100	OK split limit
chop_good_split	50	Good split limit
heuristic_segcost_rating_base	1.25	base factor for adding segmentation cost into word rating.It’s a multiplying factor, the larger the value above 1, the bigger the effect of segmentation cost.
heuristic_weight_rating	1	weight associated with char rating in combined cost of state
heuristic_weight_width	1000	weight associated with width evidence in combined cost of state
heuristic_weight_seamcut	0	weight associated with seam cut in combined cost of state
heuristic_max_char_wh_ratio	2	max char width-to-height ratio allowed in segmentation
segsearch_max_char_wh_ratio	2	Maximum character width-to-height ratio
segsearch_max_fixed_pitch_char_wh_ratio	2	Maximum character width-to-height ratio for fixed-pitch fonts
language_model_ngram_small_prob	0,000001	To avoid overly small denominators use this as the floor of the probability returned by the ngram model.
language_model_ngram_nonmatch_score	-40	Average classifier score of a non-matching unichar.
language_model_ngram_scale_factor	0.03	Strength of the character ngram model relative to the character classifier
language_model_penalty_non_freq_dict_word	0.1	Penalty for words not in the frequent word dictionary
language_model_penalty_non_dict_word	0.15	Penalty for non-dictionary words
language_model_penalty_punc	0.2	Penalty for inconsistent punctuation
language_model_penalty_case	0.1	Penalty for inconsistent case
language_model_penalty_script	0.5	Penalty for inconsistent script
language_model_penalty_chartype	0.3	Penalty for inconsistent character type
language_model_penalty_font	0	Penalty for inconsistent font
language_model_penalty_spacing	0.05	Penalty for inconsistent spacing
language_model_penalty_increment	0.01	Penalty increment
quality_rej_pc	0.08	good_quality_doc lte rejection limit
quality_blob_pc	0	good_quality_doc gte good blobs limit
quality_outline_pc	1	good_quality_doc lte outline error limit
quality_char_pc	0.95	good_quality_doc gte good char limit
test_pt_x	100000	xcoord
test_pt_y	100000	ycoord
tessedit_reject_doc_percent	65	%rej allowed before rej whole doc
tessedit_reject_block_percent	45	%rej allowed before rej whole block
tessedit_reject_row_percent	40	%rej allowed before rej whole row
tessedit_whole_wd_rej_row_percent	70	Number of row rejects in whole word rejectswhich prevents whole row rejection
tessedit_good_doc_still_rowrej_wd	1.1	rej good doc wd if more than this fraction rejected
quality_rowrej_pc	1.1	good_quality_doc gte good char limit
crunch_terrible_rating	80	crunch rating lt this
crunch_poor_garbage_cert	-9	crunch garbage cert lt this
crunch_poor_garbage_rate	60	crunch garbage rating lt this
crunch_pot_poor_rate	40	POTENTIAL crunch rating lt this
crunch_pot_poor_cert	-8	POTENTIAL crunch cert lt this
crunch_del_rating	60	POTENTIAL crunch rating lt this
crunch_del_cert	-10	POTENTIAL crunch cert lt this
crunch_del_min_ht	0.7	Del if word ht lt xht x this
crunch_del_max_ht	3	Del if word ht gt xht x this
crunch_del_min_width	3	Del if word width lt xht x this
crunch_del_high_word	1.5	Del if word gt xht x this above bl
crunch_del_low_word	0.5	Del if word gt xht x this below bl
crunch_small_outlines_size	0.6	Small if lt xht x this
fixsp_small_outlines_size	0.28	Small if lt xht x this
suspect_rating_per_ch	999.9	Dont touch bad rating limit
suspect_accept_rating	-999.9	Accept good rating limit
tessedit_lower_flip_hyphen	1.5	Aspect ratio dot/hyphen test
tessedit_upper_flip_hyphen	1.8	Aspect ratio dot/hyphen test
rej_whole_of_mostly_reject_word_fract	0.85	if >this fract
min_orientation_margin	7	Min acceptable orientation margin
tosp_old_sp_kn_th_factor	2	Factor for defining space threshold in terms of space and kern sizes
tosp_threshold_bias1	0	how far between kern and space?
tosp_threshold_bias2	0	how far between kern and space?
tosp_narrow_fraction	0.3	Fract of xheight for narrow
tosp_narrow_aspect_ratio	0.48	narrow if w/h less than this
tosp_wide_fraction	0.52	Fract of xheight for wide
tosp_wide_aspect_ratio	0	wide if w/h less than this
tosp_fuzzy_space_factor	0.6	Fract of xheight for fuzz sp
tosp_fuzzy_space_factor1	0.5	Fract of xheight for fuzz sp
tosp_fuzzy_space_factor2	0.72	Fract of xheight for fuzz sp
tosp_gap_factor	0.83	gap ratio to flip sp->kern
tosp_kern_gap_factor1	2	gap ratio to flip kern->sp
tosp_kern_gap_factor2	1.3	gap ratio to flip kern->sp
tosp_kern_gap_factor3	2.5	gap ratio to flip kern->sp
tosp_ignore_big_gaps	-1	xht multiplier
tosp_ignore_very_big_gaps	3.5	xht multiplier
tosp_rep_space	1.6	rep gap multiplier for space
tosp_enough_small_gaps	0.65	Fract of kerns reqd for isolated row stats
tosp_table_kn_sp_ratio	2.25	Min difference of kn & sp in table
tosp_table_xht_sp_ratio	0.33	Expect spaces bigger than this
tosp_table_fuzzy_kn_sp_ratio	3	Fuzzy if less than this
tosp_fuzzy_kn_fraction	0.5	New fuzzy kn alg
tosp_fuzzy_sp_fraction	0.5	New fuzzy sp alg
tosp_min_sane_kn_sp	1.5	Dont trust spaces less than this time kn
tosp_init_guess_kn_mult	2.2	Thresh guess – mult kn by this
tosp_init_guess_xht_mult	0.28	Thresh guess – mult xht by this
tosp_max_sane_kn_thresh	5	Multiplier on kn to limit thresh
tosp_flip_caution	0	Dont autoflip kn to sp when large separation
tosp_large_kerning	0.19	Limit use of xht gap with large kns
tosp_dont_fool_with_small_kerns	-1	Limit use of xht gap with odd small kns
tosp_near_lh_edge	0	Dont reduce box if the top left is non blank
tosp_silly_kn_sp_gap	0.2	Dont let sp minus kn get too small
tosp_pass_wide_fuzz_sp_to_context	0.75	How wide fuzzies need context
textord_blob_size_bigile	95	Percentile for large blobs
textord_noise_area_ratio	0.7	Fraction of bounding box for noise
textord_blob_size_smallile	20	Percentile for small blobs
textord_initialx_ile	0.75	Ile of sizes for xheight guess
textord_initialasc_ile	0.9	Ile of sizes for xheight guess
textord_noise_sizelimit	0.5	Fraction of x for big t count
textord_noise_normratio	2	Dot to norm ratio for deletion
textord_noise_syfract	0.2	xh fract height error for norm blobs
textord_noise_sxfract	0.4	xh fract width error for norm blobs
textord_noise_hfract	0.015625	Height fraction to discard outlines as speckle noise
textord_noise_rowratio	6	Dot to norm ratio for deletion
textord_blshift_maxshift	0	Max baseline shift
textord_blshift_xfraction	9.99	Min size of baseline shift