Parallel ocr evaluation for sections of screen + many refactorings

1 files changed, 14 insertions, 38 deletions

diff --git a/include/libchelleport.h b/include/libchelleport.h
index e6a074d..b69466e 100644
--- a/include/libchelleport.h
+++ b/include/libchelleport.h
@@ -1,46 +1,12 @@
+#pragma once
 #include <chrono>
+#include <iostream>
 #include <leptonica/allheaders.h>
+#include <memory>
 #include <tesseract/baseapi.h>
 #include <vector>
 
-// NOTE: Remember to update size and alignment in ocr hs module on change
-struct OCRMatch {
-  int startX, startY;
-  int endX, endY;
-  const char *text;
-};
-
-// OCR configuration
-#define CONFIDENCE_THRESHOLD 25.
-#define MIN_CHARACTER_COUNT 3
-const tesseract::PageIteratorLevel RESULT_ITER_MODE = tesseract::RIL_WORD;
-
-// Preprocessing configuration
-const float contrast = 0.3;
-const float sharpness = 0.7;
-const float scaleFactor = 1;
-const float grayscaleWeightRed = 0.114;
-const float grayscaleWeightGreen = 0.587;
-const float grayscaleWeightBlue = 0.299;
-
-extern "C" {
-OCRMatch *findWordCoordinates(const char *image_path, /* returns */ int *size);
-}
-
-tesseract::TessBaseAPI *initializeTesseract();
-
-Pix *loadImage(const char *imagePath);
-
-std::vector<OCRMatch> extractTextCoordinates(const char *imagePath);
-
-void printMatch(const OCRMatch &match);
-
-void preprocessImage(Pix **image);
-
-#define INLINE_IMAGE_PROC(process)                                             \
-  temp = process;                                                              \
-  pixDestroy(image);                                                           \
-  *image = temp;
+#include "./recognizer.h"
 
 #define MEASURE(label, stmts)                                                  \
   auto start = std::chrono::high_resolution_clock::now();                      \
@@ -49,3 +15,13 @@ void preprocessImage(Pix **image);
   auto duration =                                                              \
       std::chrono::duration_cast<std::chrono::microseconds>(end - start);      \
   std::cout << label << ": " << duration.count() / 1000.0 << " ms" << std::endl;
+
+extern "C" {
+OCRMatch *findWordCoordinates(const char *image_path, /* returns */ int *size);
+}
+
+std::vector<OCRMatch> extractTextMatches(const char *imagePath);
+
+std::vector<OCRMatch>
+runRecognizers(std::vector<std::unique_ptr<Recognizer>> &recognizers,
+               Pix *image);