Method SetupMasterShapes

deps/tesseract/classify/mastertrainer.cpp:269–305

Sets up the master_shapes_ table, which tells which fonts should stay together until they get to a leaf node classifier.

// Sets up the master_shapes_ table, which tells which fonts should stay
// together until they get to a leaf node classifier.
void MasterTrainer::SetupMasterShapes() {
  tprintf("Building master shape table\n");
  int num_fonts = samples_.NumFonts();

  ShapeTable char_shapes_begin_fragment(samples_.unicharset());
  ShapeTable char_shapes_end_fragment(samples_.unicharset());
  ShapeTable char_shapes(samples_.unicharset());
  for (int c = 0; c < samples_.charsetsize(); ++c) {
    ShapeTable shapes(samples_.unicharset());
    for (int f = 0; f < num_fonts; ++f) {
      if (samples_.NumClassSamples(f, c, true) > 0)
        shapes.AddShape(c, f);
    }
    ClusterShapes(kMinClusteredShapes, 1, kFontMergeDistance, &shapes);

    const CHAR_FRAGMENT *fragment = samples_.unicharset().get_fragment(c);

    if (fragment == NULL)
      char_shapes.AppendMasterShapes(shapes, NULL);
    else if (fragment->is_beginning())
      char_shapes_begin_fragment.AppendMasterShapes(shapes, NULL);
    else if (fragment->is_ending())
      char_shapes_end_fragment.AppendMasterShapes(shapes, NULL);
    else
      char_shapes.AppendMasterShapes(shapes, NULL);
  }
  ClusterShapes(kMinClusteredShapes, kMaxUnicharsPerCluster,
                kFontMergeDistance, &char_shapes_begin_fragment);
  char_shapes.AppendMasterShapes(char_shapes_begin_fragment, NULL);
  ClusterShapes(kMinClusteredShapes, kMaxUnicharsPerCluster,
                kFontMergeDistance, &char_shapes_end_fragment);
  char_shapes.AppendMasterShapes(char_shapes_end_fragment, NULL);
  ClusterShapes(kMinClusteredShapes, kMaxUnicharsPerCluster,
                kFontMergeDistance, &char_shapes);
  master_shapes_.AppendMasterShapes(char_shapes, NULL);
  tprintf("Master shape_table:%s\n", master_shapes_.SummaryStr().string());
}

// Adds the junk_samples_ to the main samples_ set. Junk samples are initially
// fragments and n-grams (all incorrectly segmented characters).