Add more binarization options

include/tesseract/publictypes.h

@@ -278,6 +278,17 @@ enum OcrEngineMode {
   OEM_COUNT                    // Number of OEMs
 };
 
+/** 
+ * Except when Otsu is chosen
+ * Leptonica is used for thresholding 
+ */
+enum class ThreshMethod {
+  Otsu,         // Legacy Tesseract's Otsu thresholding
+  AdaptiveOtsu,
+  TiledSauvola,
+  Count,        // Number of Thresholding methods
+};
+
 } // namespace tesseract.
 
 #endif // TESSERACT_CCSTRUCT_PUBLICTYPES_H_

src/api/baseapi.cpp

@@ -2108,20 +2108,41 @@ bool TessBaseAPI::Threshold(Pix **pix) {
     thresholder_->SetSourceYResolution(kMinCredibleResolution);
   }
   auto pageseg_mode = static_cast<PageSegMode>(static_cast<int>(tesseract_->tessedit_pageseg_mode));
-  Image im(*pix);
-  if (!thresholder_->ThresholdToPix(pageseg_mode, &im)) {
-    return false;
-  }
-  *pix = im;
+
+  Image pix_binary(*pix);
+  Image pix_grey;
+  Image pix_thresholds;
+
+  auto thresholding_method = static_cast<ThreshMethod>(static_cast<int>(tesseract_->thresholding_method));
+
+  if (thresholding_method == ThreshMethod::Otsu) {
+    if (!thresholder_->ThresholdToPix(pageseg_mode, &pix_binary)) {
+      return false;
+    }
+    *pix = pix_binary;
+
+    if (!thresholder_->IsBinary()) {
+      tesseract_->set_pix_thresholds(thresholder_->GetPixRectThresholds());
+      tesseract_->set_pix_grey(thresholder_->GetPixRectGrey());
+    } else {
+      tesseract_->set_pix_thresholds(nullptr);
+      tesseract_->set_pix_grey(nullptr);
+    }
+  } else {
+    auto [ok, pix_grey, pix_binary, pix_thresholds] = thresholder_->Threshold(thresholding_method);
+
+    if (!ok) {
+      return false;
+    }
+    *pix = pix_binary;
+
+    tesseract_->set_pix_thresholds(pix_thresholds);
+    tesseract_->set_pix_grey(pix_grey); 
+  } 
+
   thresholder_->GetImageSizes(&rect_left_, &rect_top_, &rect_width_, &rect_height_, &image_width_,
                               &image_height_);
-  if (!thresholder_->IsBinary()) {
-    tesseract_->set_pix_thresholds(thresholder_->GetPixRectThresholds());
-    tesseract_->set_pix_grey(thresholder_->GetPixRectGrey());
-  } else {
-    tesseract_->set_pix_thresholds(nullptr);
-    tesseract_->set_pix_grey(nullptr);
-  }
+
   // Set the internal resolution that is used for layout parameters from the
   // estimated resolution, rather than the image resolution, which may be
   // fabricated, but we will use the image resolution, if there is one, to

src/ccmain/tesseractclass.cpp

@@ -74,6 +74,11 @@ Tesseract::Tesseract()
                "11=sparse_text, 12=sparse_text+osd, 13=raw_line"
                " (Values from PageSegMode enum in tesseract/publictypes.h)",
                this->params())
+    , INT_MEMBER(thresholding_method,
+                 static_cast<int>(tesseract::ThreshMethod::Otsu),                
+"Thresholding "
+                 "method: 0 = Otsu, 1 = Adaptive Otsu, 2 = Sauvola",
+                 this->params())
     , INT_INIT_MEMBER(tessedit_ocr_engine_mode, tesseract::OEM_DEFAULT,
                       "Which OCR engine(s) to run (Tesseract, LSTM, both)."
                       " Defaults to loading and running the most accurate"

src/ccmain/tesseractclass.h

@@ -762,6 +762,9 @@ class TESS_API Tesseract : public Wordrec {
             "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 tesseract/publictypes.h)");
+  INT_VAR_H(thresholding_method,
+            static_cast<int>(tesseract::ThreshMethod::Otsu), "Thresholding "
+            "method: 0 = Otsu, 1 = Adaptive Otsu, 2 = Sauvola");
   INT_VAR_H(tessedit_ocr_engine_mode, tesseract::OEM_DEFAULT,
             "Which OCR engine(s) to run (Tesseract, LSTM, both). Defaults"
             " to loading and running the most accurate available.");

src/ccmain/thresholder.cpp

@@ -20,6 +20,7 @@
 
 #include <cstdint> // for uint32_t
 #include <cstring>
+#include <tuple>
 
 #include "otsuthr.h"
 #include "thresholder.h"
@@ -184,6 +185,45 @@ void ImageThresholder::SetImage(const Image pix) {
   Init();
 }
 
+std::tuple<bool, Image, Image, Image> ImageThresholder::Threshold(
+                                                         ThreshMethod method) {
+  Image pix_grey = nullptr;
+  Image pix_binary = nullptr;
+  Image pix_thresholds = nullptr;
+
+  if (image_width_ > INT16_MAX || image_height_ > INT16_MAX) {
+    tprintf("Image too large: (%d, %d)\n", image_width_, image_height_);
+    return std::make_tuple(false, nullptr, nullptr, nullptr);
+  }
+
+  if (pix_channels_ == 0) {
+    // We have a binary image, but it still has to be copied, as this API
+    // allows the caller to modify the output.
+    Image original = GetPixRect();
+    pix_binary = original.copy();
+    original.destroy();
+    return std::make_tuple(false, nullptr, pix_binary, nullptr);
+  }
+
+  pix_grey = GetPixRectGrey();
+
+  if (method == ThreshMethod::Otsu || method >= ThreshMethod::Count) {
+    method = ThreshMethod::AdaptiveOtsu;
+  }
+
+  int r;
+  if (method == ThreshMethod::AdaptiveOtsu) {
+    r = pixOtsuAdaptiveThreshold(pix_grey, 300, 300, 0, 0, 0.1, 
+                                 pix_thresholds.a(), pix_binary.a());
+  } else if (method == ThreshMethod::TiledSauvola) {
+    r = pixSauvolaBinarizeTiled(pix_grey, 25, 0.40, 300, 300, pix_thresholds.a(), 
+                                pix_binary.a());
+  }
+
+  bool ok = r == 0 ? true : false;
+  return std::make_tuple(ok, pix_grey, pix_binary, pix_thresholds);
+}
+
 // Threshold the source image as efficiently as possible to the output Pix.
 // Creates a Pix and sets pix to point to the resulting pointer.
 // Caller must use pixDestroy to free the created Pix.

src/ccmain/thresholder.h

@@ -121,6 +121,9 @@ class TESS_API ImageThresholder {
   /// Returns false on error.
   virtual bool ThresholdToPix(PageSegMode pageseg_mode, Image *pix);
 
+  virtual std::tuple<bool, Image, Image, Image> Threshold(
+                                                          ThreshMethod method);
+
   // Gets a pix that contains an 8 bit threshold value at each pixel. The
   // returned pix may be an integer reduction of the binary image such that
   // the scale factor may be inferred from the ratio of the sizes, even down

src/ccstruct/image.h

@@ -33,6 +33,8 @@ class TESS_API Image {
   // service
   operator Pix *() const { return pix_; }
   Pix *operator->() const { return pix_; }
+  Pix **a() { return &pix_; }
+
 
   // api
   Image clone() const; // increases refcount