From 70e3920556496e5fecb5fedddf1067b2522fcac7 Mon Sep 17 00:00:00 2001
From: Akshay Nair <phenax5@gmail.com>
Date: Tue, 24 Dec 2024 18:51:17 +0530
Subject: Add setup for ocr with tesseract

---
 cpp/libchelleport.cpp | 74 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 74 insertions(+)
 create mode 100644 cpp/libchelleport.cpp

(limited to 'cpp')

diff --git a/cpp/libchelleport.cpp b/cpp/libchelleport.cpp
new file mode 100644
index 0000000..8f2e9f2
--- /dev/null
+++ b/cpp/libchelleport.cpp
@@ -0,0 +1,74 @@
+#include <algorithm>
+#include <cstdio>
+#include <cstdlib>
+#include <cstring>
+#include <iostream>
+#include <leptonica/allheaders.h>
+#include <tesseract/baseapi.h>
+#include <tesseract/publictypes.h>
+#include <vector>
+
+#include "../include/libchelleport.h"
+
+std::vector<OCRMatch> extractTextCoordinates(const char *imagePath);
+
+#define CONFIDENCE_THRESHOLD 30.
+
+OCRMatch *findWordCoordinates(const char *image_path, int *size) {
+  auto boxes = extractTextCoordinates(image_path);
+  static OCRMatch *ptr = new OCRMatch[boxes.size()];
+  std::copy(boxes.begin(), boxes.end(), ptr);
+
+  // for (const auto &box : boxes) {
+  //   std::cout << box.text << "\n\n";
+  //   std::cout << "Text: " << box.text << "\nPosition: (" << box.startX << ","
+  //             << box.startY << ") -> (" << box.endX << "," << box.endY << ")"
+  //             << "\n\n";
+  // }
+
+  *size = boxes.size();
+  return ptr;
+}
+
+std::vector<OCRMatch> extractTextCoordinates(const char *imagePath) {
+  std::vector<OCRMatch> results;
+  tesseract::TessBaseAPI *tesseract = new tesseract::TessBaseAPI();
+
+  if (tesseract->Init(nullptr, "eng")) {
+    std::cerr << "Could not initialize tesseract." << std::endl;
+    return results;
+  }
+
+  Pix *image = pixRead(imagePath);
+  if (!image) {
+    std::cerr << "Could not load image " << imagePath << std::endl;
+    return results;
+  }
+
+  tesseract->SetImage(image);
+  tesseract->Recognize(0);
+
+  tesseract::ResultIterator *iterator = tesseract->GetIterator();
+  tesseract::PageIteratorLevel level = tesseract::RIL_TEXTLINE;
+
+  if (iterator != 0) {
+    do {
+      float conf = iterator->Confidence(level);
+      const char *word = iterator->GetUTF8Text(level);
+      int x1, y1, x2, y2;
+      iterator->BoundingBox(level, &x1, &y1, &x2, &y2);
+
+      if (conf > CONFIDENCE_THRESHOLD && word != nullptr && strlen(word) >= 2) {
+        OCRMatch box{x1, y1, x2, y2, word};
+        results.push_back(box);
+      }
+    } while (iterator->Next(level));
+  }
+
+  delete iterator;
+  tesseract->End();
+  delete tesseract;
+  pixDestroy(&image);
+
+  return results;
+}
-- 
cgit v1.3.1