Rewrite of DENORM

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@538 d0cd1f9f-072b-0410-8dd7-cf729c803f20

ccmain/Makefile.in

@@ -251,6 +251,7 @@ libdir = @libdir@
 libexecdir = @libexecdir@
 localedir = @localedir@
 localstatedir = @localstatedir@
+lt_ECHO = @lt_ECHO@
 mandir = @mandir@
 mkdir_p = @mkdir_p@
 oldincludedir = @oldincludedir@

ccmain/control.cpp

@@ -345,7 +345,7 @@ void Tesseract::recog_all_words(PAGE_RES* page_res,
   set_global_loc_code(LOC_FUZZY_SPACE);
 
   if (!tessedit_test_adaption && tessedit_fix_fuzzy_spaces
-      && !tessedit_word_for_word)
+      && !tessedit_word_for_word && !right_to_left())
     fix_fuzzy_spaces(monitor, stats_.word_count, page_res);
 
   // ****************** Pass 4 *******************
@@ -455,6 +455,10 @@ void Tesseract::recog_all_words(PAGE_RES* page_res,
   PageSegMode pageseg_mode = static_cast<PageSegMode>(
       static_cast<int>(tessedit_pageseg_mode));
   textord_.CleanupSingleRowResult(pageseg_mode, page_res);
+
+  if (monitor != NULL) {
+    monitor->progress = 100;
+  }
 }
 
 

ccmain/fixxht.cpp

@@ -154,11 +154,11 @@ float Tesseract::ComputeCompatibleXheight(WERD_RES *word_res) {
   if (debug_x_ht_level >= 20) {
     tprintf("Median xht=%f\n", new_xht);
     tprintf("Mode20:A: New x-height = %f (norm), %f (orig)\n",
-            new_xht, new_xht / word_res->denorm.scale());
+            new_xht, new_xht / word_res->denorm.y_scale());
   }
   // The xheight must change by at least x_ht_min_change to be used.
   if (fabs(new_xht - kBlnXHeight) >= x_ht_min_change)
-    return new_xht / word_res->denorm.scale();
+    return new_xht / word_res->denorm.y_scale();
   else
     return 0.0f;
 }

ccmain/osdetect.cpp

@@ -280,57 +280,42 @@ int os_detect_blobs(BLOBNBOX_CLIST* blob_list, OSResults* osr,
 bool os_detect_blob(BLOBNBOX* bbox, OrientationDetector* o,
                     ScriptDetector* s, OSResults* osr,
                     tesseract::Tesseract* tess) {
-  C_BLOB*   blob = bbox->cblob();
-  TBOX      box = blob->bounding_box();
-
-  int       x_mid = (box.left() + box.right()) / 2.0f;
-  int       y_mid = (box.bottom() + box.top()) / 2.0f;
-
-  PBLOB     pblob(blob);
-
+  tess->tess_cn_matching.set_value(true); // turn it on
+  tess->tess_bn_matching.set_value(false);
+  C_BLOB* blob = bbox->cblob();
+  TBLOB* tblob = TBLOB::PolygonalCopy(blob);
+  TBOX box = tblob->bounding_box();
+  FCOORD current_rotation(1.0f, 0.0f);
+  FCOORD rotation90(0.0f, 1.0f);
   BLOB_CHOICE_LIST ratings[4];
   // Test the 4 orientations
   for (int i = 0; i < 4; ++i) {
-    // normalize the blob
+    // Normalize the blob. Set the origin to the place we want to be the
+    // bottom-middle after rotation.
+    // Scaling is to make the rotated height the x-height.
     float scaling = static_cast<float>(kBlnXHeight) / box.height();
-    DENORM denorm(x_mid, scaling, 0.0, box.bottom(), 0, NULL, false, NULL);
-    pblob.move(FCOORD(-x_mid, -box.bottom()));
-    pblob.scale(scaling);
-    pblob.move(FCOORD(0.0f, kBlnBaselineOffset));
-
-    {
-      // List of choices given by the classifier
-      tess->tess_cn_matching.set_value(true); // turn it on
-      tess->tess_bn_matching.set_value(false);
-      // Convert blob
-      TBLOB* tessblob = make_tess_blob(&pblob);
-      // Classify
-      tess->set_denorm(&denorm);
-      tess->AdaptiveClassifier(tessblob, ratings + i, NULL);
-      delete tessblob;
+    float x_origin = (box.left() + box.right()) / 2.0f;
+    float y_origin = (box.bottom() + box.top()) / 2.0f;
+    if (i == 0 || i == 2) {
+      // Rotation is 0 or 180.
+      y_origin = i == 0 ? box.bottom() : box.top();
+    } else {
+      // Rotation is 90 or 270.
+      scaling = static_cast<float>(kBlnXHeight) / box.width();
+      x_origin = i == 1 ? box.left() : box.right();
     }
-    // undo normalize
-    pblob.move(FCOORD(0.0f, -kBlnBaselineOffset));
-    pblob.scale(1.0f / scaling);
-    pblob.move(FCOORD(x_mid, box.bottom()));
-
-    // center the blob
-    pblob.move(FCOORD(-x_mid, -y_mid));
-
-    // TODO(rays) Although we should now get the correct image coords with
-    // the DENORM, there is nothing to tell the classifier to rotate the
-    // image or to actually rotate the image for it.
-    // Rotate it
-    pblob.rotate();
-
-    // Re-compute the mid
-    box = pblob.bounding_box();
-    x_mid = (box.left() + box.right()) / 2;
-    y_mid = (box.top() + box.bottom()) / 2;
-
-    // re-center in the new mid
-    pblob.move(FCOORD(x_mid, y_mid));
+    DENORM denorm;
+    denorm.SetupNormalization(NULL, NULL, &current_rotation, NULL, NULL, 0,
+                              x_origin, y_origin, scaling, scaling,
+                              0.0f, static_cast<float>(kBlnBaselineOffset));
+    TBLOB* rotated_blob = new TBLOB(*tblob);
+    rotated_blob->Normalize(denorm);
+    tess->set_denorm(&denorm);
+    tess->AdaptiveClassifier(rotated_blob, ratings + i, NULL);
+    delete rotated_blob;
+    current_rotation.rotate(rotation90);
   }
+  delete tblob;
 
   bool stop = o->detect_blob(ratings);
   s->detect_blob(ratings);

ccmain/reject.cpp

@@ -220,7 +220,7 @@ void Tesseract::make_reject_map(      //make rej map for wd //detailed results
     and the whole of any words which are very small
   */
   else if (tessedit_reject_mode == 5) {
-    if (kBlnXHeight / word->denorm.scale () <= min_sane_x_ht_pixels)
+    if (kBlnXHeight / word->denorm.y_scale() <= min_sane_x_ht_pixels)
       word->reject_map.rej_word_small_xht ();
     else {
       one_ell_conflict(word, TRUE);
@@ -820,7 +820,7 @@ void Tesseract::flip_hyphens(WERD_RES *word_res) {
     else
       next_left = blob->next->bounding_box().left();
     // Dont touch small or touching blobs - it is too dangerous.
-    if ((out_box.width() > 8 * word_res->denorm.scale()) &&
+    if ((out_box.width() > 8 * word_res->denorm.x_scale()) &&
         (out_box.left() > prev_right) && (out_box.right() < next_left)) {
       aspect_ratio = out_box.width() / (float) out_box.height();
       if (unicharset.eq(best_choice->unichar_id(i), ".")) {

ccmain/tessedit.cpp

@@ -109,8 +109,10 @@ bool Tesseract::init_tesseract_lang_data(
 
   // Initialize TessdataManager.
   STRING tessdata_path = language_data_path_prefix + kTrainedDataSuffix;
-  tessdata_manager.Init(tessdata_path.string(),
-                        tessdata_manager_debug_level);
+  if (!tessdata_manager.Init(tessdata_path.string(),
+                             tessdata_manager_debug_level)) {
+    return false;
+  }
 
   // If a language specific config file (lang.config) exists, load it in.
   if (tessdata_manager.SeekToStart(TESSDATA_LANG_CONFIG)) {

ccmain/tesseractclass.cpp

@@ -75,14 +75,9 @@ Tesseract::Tesseract()
                 " a character composed form fragments", this->params()),
     BOOL_MEMBER(tessedit_adaption_debug, false, "Generate and print debug"
                 " information for adaption", this->params()),
-    BOOL_MEMBER(applybox_rebalance, TRUE, "Drop dead", this->params()),
     INT_MEMBER(applybox_debug, 1, "Debug level", this->params()),
     INT_MEMBER(applybox_page, 0,
                "Page number to apply boxes from", this->params()),
-    STRING_MEMBER(applybox_test_exclusions, "",
-                  "Chars ignored for testing", this->params()),
-    double_MEMBER(applybox_error_band, 0.15,
-                  "Err band as fract of xht", this->params()),
     STRING_MEMBER(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"
@@ -94,10 +89,6 @@ Tesseract::Tesseract()
     BOOL_MEMBER(applybox_learn_ngrams_mode, false, "Each bounding box"
                 " is assumed to contain ngrams. Only learn the ngrams"
                 " whose outlines overlap horizontally.", this->params()),
-    BOOL_MEMBER(tessedit_print_text, false,
-                "Write text to stdout", this->params()),
-    BOOL_MEMBER(tessedit_draw_words, false,
-                "Draw source words", this->params()),
     BOOL_MEMBER(tessedit_draw_outwords, false,
                 "Draw output words", this->params()),
     BOOL_MEMBER(tessedit_training_tess, false,
@@ -277,12 +268,6 @@ Tesseract::Tesseract()
                "Min change in xht before actually trying it", this->params()),
     BOOL_MEMBER(tessedit_write_block_separators, false,
                 "Write block separators in output", this->params()),
-    BOOL_MEMBER(tessedit_write_raw_output, false,
-                "Write raw stuff to name.raw", this->params()),
-    BOOL_MEMBER(tessedit_write_output, false,
-                "Write text to name.txt", this->params()),
-    BOOL_MEMBER(tessedit_write_ratings, false,
-                "Return ratings in IPEOCRAPI data", this->params()),
     BOOL_MEMBER(tessedit_write_rep_codes, false,
                 "Write repetition char code", this->params()),
     BOOL_MEMBER(tessedit_write_unlv, false,
@@ -351,18 +336,12 @@ Tesseract::Tesseract()
                "Reject any x-ht lt or eq than this", this->params()),
     BOOL_MEMBER(tessedit_create_boxfile, false,
                 "Output text with boxes", this->params()),
-    BOOL_MEMBER(tessedit_read_image, true,
-                "Ensure the image is read", this->params()),
-    INT_MEMBER(tessedit_serial_unlv, 0, "0->Whole page, 1->serial"
-               " no adapt, 2->serial with adapt", this->params()),
     INT_MEMBER(tessedit_page_number, -1, "-1 -> All pages"
                " , else specifc page to process", this->params()),
     BOOL_MEMBER(tessedit_write_images, false,
                 "Capture the image from the IPE", this->params()),
     BOOL_MEMBER(interactive_mode, false, "Run interactively?", this->params()),
     STRING_MEMBER(file_type, ".tif", "Filename extension", this->params()),
-    INT_MEMBER(testedit_match_debug, 0,
-               "Integer match debug ctrl", this->params()),
     BOOL_MEMBER(tessedit_override_permuter, true,
                 "According to dict_word", this->params()),
     INT_INIT_MEMBER(tessdata_manager_debug_level, 0, "Debug level for"

ccmain/tesseractclass.h

@@ -562,11 +562,8 @@ class Tesseract : public Wordrec {
              " a character composed form fragments");
   BOOL_VAR_H(tessedit_adaption_debug, false,
              "Generate and print debug information for adaption");
-  BOOL_VAR_H(applybox_rebalance, true, "Drop dead");
   INT_VAR_H(applybox_debug, 1, "Debug level");
   INT_VAR_H(applybox_page, 0, "Page number to apply boxes from");
-  STRING_VAR_H(applybox_test_exclusions, "", "Chars ignored for testing");
-  double_VAR_H(applybox_error_band, 0.15, "Err band as fract of xht");
   STRING_VAR_H(applybox_exposure_pattern, ".exp",
                "Exposure value follows this pattern in the image"
                " filename. The name of the image files are expected"
@@ -578,8 +575,6 @@ class Tesseract : public Wordrec {
   BOOL_VAR_H(applybox_learn_ngrams_mode, false,
              "Each bounding box is assumed to contain ngrams. Only"
              " learn the ngrams whose outlines overlap horizontally.");
-  BOOL_VAR_H(tessedit_print_text, false, "Write text to stdout");
-  BOOL_VAR_H(tessedit_draw_words, false, "Draw source words");
   BOOL_VAR_H(tessedit_draw_outwords, false, "Draw output words");
   BOOL_VAR_H(tessedit_training_tess, false, "Call Tess to learn blobs");
   BOOL_VAR_H(tessedit_dump_choices, false, "Dump char choices");
@@ -706,11 +701,6 @@ class Tesseract : public Wordrec {
   INT_VAR_H(x_ht_min_change, 8, "Min change in xht before actually trying it");
   BOOL_VAR_H(tessedit_write_block_separators, false,
              "Write block separators in output");
-  BOOL_VAR_H(tessedit_write_raw_output, false,
-             "Write raw stuff to name.raw");
-  BOOL_VAR_H(tessedit_write_output, false, "Write text to name.txt");
-  BOOL_VAR_H(tessedit_write_ratings, false,
-             "Return ratings in IPEOCRAPI data");
   BOOL_VAR_H(tessedit_write_rep_codes, false,
              "Write repetition char code");
   BOOL_VAR_H(tessedit_write_unlv, false, "Write .unlv output file");
@@ -755,15 +745,11 @@ class Tesseract : public Wordrec {
   STRING_VAR_H(conflict_set_I_l_1, "Il1[]", "Il1 conflict set");
   INT_VAR_H(min_sane_x_ht_pixels, 8, "Reject any x-ht lt or eq than this");
   BOOL_VAR_H(tessedit_create_boxfile, false, "Output text with boxes");
-  BOOL_VAR_H(tessedit_read_image, true, "Ensure the image is read");
-  INT_VAR_H(tessedit_serial_unlv, 0,
-            "0->Whole page, 1->serial no adapt, 2->serial with adapt");
   INT_VAR_H(tessedit_page_number, -1,
             "-1 -> All pages, else specifc page to process");
   BOOL_VAR_H(tessedit_write_images, false, "Capture the image from the IPE");
   BOOL_VAR_H(interactive_mode, false, "Run interactively?");
   STRING_VAR_H(file_type, ".tif", "Filename extension");
-  INT_VAR_H(testedit_match_debug, 0, "Integer match debug ctrl");
   BOOL_VAR_H(tessedit_override_permuter, true, "According to dict_word");
   INT_VAR_H(tessdata_manager_debug_level, 0,
             "Debug level for TessdataManager functions.");

ccmain/werdit.cpp

-
 /**********************************************************************
  * File:        werdit.cpp  (Formerly wordit.c)
  * Description: An iterator for passing over all the words in a document.