Add configurable variables to control thresholding (#3577)

tesseract-ocr · Sep 29, 2021 · 0cb9c40 · 0cb9c40
1 parent c4ad9b7
commit 0cb9c40
Show file tree

Hide file tree

Showing 5 changed files with 64 additions and 15 deletions.
diff --git a/src/api/baseapi.cpp b/src/api/baseapi.cpp
@@ -2072,7 +2072,7 @@ bool TessBaseAPI::Threshold(Pix **pix) {
       tesseract_->set_pix_grey(nullptr);
     }
   } else {
-    auto [ok, pix_grey, pix_binary, pix_thresholds] = thresholder_->Threshold(thresholding_method);
+    auto [ok, pix_grey, pix_binary, pix_thresholds] = thresholder_->Threshold(this, thresholding_method);
 
     if (!ok) {
       return false;

diff --git a/src/ccmain/tesseractclass.cpp b/src/ccmain/tesseractclass.cpp
@@ -76,8 +76,34 @@ Tesseract::Tesseract()
                this->params())
     , INT_MEMBER(thresholding_method,
                  static_cast<int>(tesseract::ThresholdMethod::Otsu),
-                 "Thresholding "
-                 "method: 0 = Otsu, 1 = Adaptive Otsu, 2 = Sauvola",
+                 "Thresholding method: 0 = Otsu, 1 = Adaptive Otsu, 2 = "
+                 "Sauvola",
+                 this->params())
+    , INT_MEMBER(thresholding_window_size, 51,
+                 "Window size for measuring local statistics. "
+                 "This parameter is used by the Sauvola thresolding method",
+                 this->params())
+    , double_MEMBER(thresholding_kfactor, 0.34,
+                    "Factor for reducing threshold due to variance. "
+                    "This parameter is used by the Sauvola thresolding method. "
+                    "Must be >= 0",
+                 this->params())
+    , INT_MEMBER(thresholding_tile_size, 300,
+                 "Desired tile size. Actual size may vary. Must be >= 16. "
+                 "This parameter is used by the Adaptive Otsu thresolding "
+                 "method",
+                 this->params())
+    , INT_MEMBER(thresholding_smooth_size, 0,
+                 "Size of convolution kernel applied to threshold array. "
+                 "This parameter is used by the Adaptive Otsu thresolding "
+                 "method. "
+                 "Use 0 for no smoothing",
+                 this->params())
+    , double_MEMBER(thresholding_score_fraction, 0.1,
+                 "Fraction of the max Otsu score. "
+                 "This parameter is used by the Adaptive Otsu thresolding "
+                 "method. "
+                 "Typically 0.1. Use 0.0 for standard Otsu",
                  this->params())
     , INT_INIT_MEMBER(tessedit_ocr_engine_mode, tesseract::OEM_DEFAULT,
                       "Which OCR engine(s) to run (Tesseract, LSTM, both)."

diff --git a/src/ccmain/tesseractclass.h b/src/ccmain/tesseractclass.h
@@ -757,6 +757,11 @@ class TESS_API Tesseract : public Wordrec {
   BOOL_VAR_H(tessedit_do_invert);
   INT_VAR_H(tessedit_pageseg_mode);
   INT_VAR_H(thresholding_method);
+  INT_VAR_H(thresholding_window_size);
+  double_VAR_H(thresholding_kfactor);
+  INT_VAR_H(thresholding_tile_size);
+  INT_VAR_H(thresholding_smooth_size);
+  double_VAR_H(thresholding_score_fraction);
   INT_VAR_H(tessedit_ocr_engine_mode);
   STRING_VAR_H(tessedit_char_blacklist);
   STRING_VAR_H(tessedit_char_whitelist);

diff --git a/src/ccmain/thresholder.cpp b/src/ccmain/thresholder.cpp
@@ -25,6 +25,7 @@
 #endif
 
 #include <allheaders.h>
+#include <tesseract/baseapi.h> // for api->GetIntVariable()
 
 #include <cstdint> // for uint32_t
 #include <cstring>
@@ -186,7 +187,8 @@ void ImageThresholder::SetImage(const Image pix) {
 }
 
 std::tuple<bool, Image, Image, Image> ImageThresholder::Threshold(
-                                                         ThresholdMethod method) {
+                                                      TessBaseAPI *api,
+                                                      ThresholdMethod method) {
   Image pix_binary = nullptr;
   Image pix_thresholds = nullptr;
 
@@ -203,9 +205,10 @@ std::tuple<bool, Image, Image, Image> ImageThresholder::Threshold(
 
   int r;
   if (method == ThresholdMethod::Sauvola) {
-    // TODO: Convert this constant to config var
-    // window half-width for measuring local statistics
-    constexpr l_int32 whsize = 25;
+    bool b;
+    int window_size;
+    b = api->GetIntVariable("thresholding_window_size", &window_size);
+    int half_window_size = window_size / 2;
     // factor for image division into tiles; >= 1
     l_int32 nx, ny;
 //  // tiles size will be approx. 250 x 250 pixels
@@ -215,19 +218,32 @@ std::tuple<bool, Image, Image, Image> ImageThresholder::Threshold(
     ny = std::max(1, (pix_h + 125) / 250);
     auto xrat = pix_w / nx;
     auto yrat = pix_h / ny;
-    if (xrat < whsize + 2) {
-      nx = pix_w / (whsize + 2);
+    if (xrat < half_window_size + 2) {
+      nx = pix_w / (half_window_size + 2);
     }
-    if (yrat < whsize + 2) {
-      ny = pix_h / (whsize + 2);
+    if (yrat < half_window_size + 2) {
+      ny = pix_h / (half_window_size + 2);
     }
 
-    r = pixSauvolaBinarizeTiled(pix_grey, whsize, 0.40, nx, ny,
+    double kfactor;
+    b = api->GetDoubleVariable("thresholding_kfactor", &kfactor);
+    r = pixSauvolaBinarizeTiled(pix_grey, half_window_size, kfactor, nx, ny,
                                (PIX**)pix_thresholds,
                                 (PIX**)pix_binary);
   } else { // if (method == ThresholdMethod::AdaptiveOtsu)
-    r = pixOtsuAdaptiveThreshold(pix_grey, 300, 300, 0, 0, 0.1,
-                                 (PIX**)pix_thresholds, (PIX**)pix_binary);
+    bool b;
+    int tile_size;
+    b = api->GetIntVariable("thresholding_tile_size", &tile_size);
+    int smooth_size;
+    b = api->GetIntVariable("thresholding_smooth_size", &smooth_size);
+    int half_smooth_size = smooth_size / 2;
+    double score_fraction;
+    b = api->GetDoubleVariable("thresholding_score_fraction", &score_fraction);
+    r = pixOtsuAdaptiveThreshold(pix_grey, tile_size, tile_size,
+                                 half_smooth_size, half_smooth_size,
+                                 score_fraction, 
+                                 (PIX**)pix_thresholds,
+                                 (PIX**)pix_binary);
   }
 
   bool ok = (r == 0);

diff --git a/src/ccmain/thresholder.h b/src/ccmain/thresholder.h
@@ -28,6 +28,8 @@ struct Pix;
 
 namespace tesseract {
 
+class TessBaseAPI;
+
 /// Base class for all tesseract image thresholding classes.
 /// Specific classes can add new thresholding methods by
 /// overriding ThresholdToPix.
@@ -121,7 +123,7 @@ class TESS_API ImageThresholder {
   /// Returns false on error.
   virtual bool ThresholdToPix(Image *pix);
 
-  virtual std::tuple<bool, Image, Image, Image> Threshold(
+  virtual std::tuple<bool, Image, Image, Image> Threshold(TessBaseAPI *api,
                                                           ThresholdMethod method);
 
   // Gets a pix that contains an 8 bit threshold value at each pixel. The