Code cleanup and color recognition base

src/main/java/es/yonatan/tfg/Main.java

 package es.yonatan.tfg;
 
 import es.yonatan.tfg.boxing.HasBoundingBox;
-import es.yonatan.tfg.boxing.Sentence;
+import es.yonatan.tfg.boxing.UnitGroup;
 import es.yonatan.tfg.boxing.Unit;
+import es.yonatan.tfg.recognizer.ColorRecognizer;
 import es.yonatan.tfg.recognizer.TesseractRecognizer;
 import es.yonatan.tfg.util.HashGrid2D;
 import net.sourceforge.tess4j.util.ImageHelper;
@@ -13,15 +14,19 @@ import java.awt.*;
 import java.awt.image.BufferedImage;
 import java.io.File;
 import java.io.IOException;
-import java.util.Comparator;
 import java.util.HashSet;
 
 public class Main {
 	public static void main(String[] args) throws IOException {
+		var file = new File("C:/Users/Yonatan/Desktop/04.jpg");
+
 		var time = System.currentTimeMillis();
-		var originalImage = ImageIO.read(new File("C:/Users/Yonatan/Desktop/04.jpg"));
+		//TODO: 10/04/2023 transform originalImage for best Tesseract OCR results: grayscale, binarize, noise reduction, etc.
+		var originalImage = ImageIO.read(file);
 		var image = ImageHelper.convertImageToGrayscale(originalImage);
-		//image = ImageHelper.convertImageToBinary(originalImage);
+		image = ImageHelper.convertImageToBinary(image);
+
+		System.out.println(new ColorRecognizer().getColorAt(file, 0, 0, 10, 10));
 
 		var tess = new TesseractRecognizer.Builder()
 				.setDataPath(new File("C:/Proyectos/TFG/src/main/resources/tessdata/"))
@@ -32,24 +37,23 @@ public class Main {
 
 		var words = tess.recognize(image, 80);
 
-		System.out.println(words);
-
 		var grid = new HashGrid2D<>(image.getWidth(), image.getHeight(), 4, 4);
 
 		words.forEach(grid::insert);
 
-//		merge(grid);
-		merge(grid);
+		int mergeCount = 1;
+
+		for (int i = 0; i < mergeCount; i++) {
 			merge(grid);
+		}
 
+		// flatten the grid to get a list of all the words
 		var sentences = grid.flatten().stream()
-				//.filter(element -> element instanceof Sentence)
 				.toList();
 
-		sentences.stream().sorted(Comparator.comparingInt(value -> value.bounds().y())).forEach(System.out::println);
+		// basic GUI to render the bounding boxes resulting of the OCR recognition as well as testing the word merge algorithm
 
 		var scale = 1.5;
-
 		var window = new JFrame();
 		window.setUndecorated(true);
 		window.setSize(new Dimension((int) (originalImage.getWidth() / scale), (int) (originalImage.getHeight() / scale)));
@@ -98,28 +102,28 @@ public class Main {
 
 			grid.remove(word);
 
-			Sentence sentence;
-			if (word instanceof Sentence) sentence = (Sentence) word;
+			UnitGroup unitGroup;
+			if (word instanceof UnitGroup) unitGroup = (UnitGroup) word;
 			else {
 
-				sentence = new Sentence();
-				sentence.addWord((Unit) word, false);
+				unitGroup = new UnitGroup();
+				unitGroup.addWord((Unit) word, false);
 			}
 
 			for (HasBoundingBox collider : colliders) {
 				grid.remove(collider);
 
-				if (collider instanceof Sentence) {
-					sentence = Sentence.merge(sentence, (Sentence) collider);
+				if (collider instanceof UnitGroup) {
+					unitGroup = UnitGroup.merge(unitGroup, (UnitGroup) collider);
 				} else {
 					if (checked.contains(collider)) continue;
 					checked.add(collider);
 
-					sentence.addWord((Unit) collider, false);
+					unitGroup.addWord((Unit) collider, false);
 				}
 			}
-			sentence.recomputeOrder();
-			grid.insert(sentence);
+			unitGroup.recomputeOrder();
+			grid.insert(unitGroup);
 		}
 	}
 }
\ No newline at end of file

src/main/java/es/yonatan/tfg/boxing/Unit.java

 package es.yonatan.tfg.boxing;
 
 import es.yonatan.tfg.util.BoundingBox;
-import es.yonatan.tfg.util.ColorData;
 import org.jetbrains.annotations.NotNull;
-import org.jetbrains.annotations.Nullable;
 
-public record Unit(@NotNull String text, float confidence, @NotNull BoundingBox bounds, @Nullable ColorData colorData) implements HasBoundingBox {
+public record Unit(@NotNull String text, float confidence, @NotNull BoundingBox bounds) implements HasBoundingBox {
 
-	public static final Unit NEXT_IMAGE_UNIT = new Unit("next_image", 100, new BoundingBox(-1,-1,-1,-1), null);
+	public static final Unit NEXT_IMAGE_UNIT = new Unit("next_image", 100, new BoundingBox(-1,-1,-1,-1));
 
 	@Override
 	public String toString() {

src/main/java/es/yonatan/tfg/boxing/Sentence.java → src/main/java/es/yonatan/tfg/boxing/UnitGroup.java

@@ -9,25 +9,25 @@ import java.util.Comparator;
 import java.util.LinkedList;
 import java.util.List;
 
-public class Sentence implements HasBoundingBox {
+public class UnitGroup implements HasBoundingBox {
 
 	private final @NotNull LinkedList<Unit> words;
 
 	private @Nullable BoundingBox bounds;
 
-	public Sentence() {
+	public UnitGroup() {
 		words = new LinkedList<>();
 	}
 
-	public static Sentence merge(Sentence... sentences) {
-		return merge(List.of(sentences));
+	public static UnitGroup merge(UnitGroup... unitGroups) {
+		return merge(List.of(unitGroups));
 	}
 
-	public static Sentence merge(Collection<Sentence> sentences) {
-		var merged = new Sentence();
+	public static UnitGroup merge(Collection<UnitGroup> unitGroups) {
+		var merged = new UnitGroup();
 
-		for (Sentence sentence : sentences) {
-			merged.addWords(sentence.words, false);
+		for (UnitGroup unitGroup : unitGroups) {
+			merged.addWords(unitGroup.words, false);
 		}
 		merged.recomputeOrder();
 		return merged;

src/main/java/es/yonatan/tfg/recognizer/ColorRecognizer.java

+package es.yonatan.tfg.recognizer;
+
+import java.awt.*;
+import java.io.File;
+
+public class ColorRecognizer implements IColorRecognizer {
+
+	@Override
+	public Color getColorAt(File image, int u, int v, int width, int height) {
+		// TODO: 4/04/2023 get primary color of image using leptonica library lept4j
+	return Color.WHITE;
+	}
+}

src/main/java/es/yonatan/tfg/recognizer/IColorRecognizer.java

+package es.yonatan.tfg.recognizer;
+
+import java.awt.*;
+import java.io.File;
+
+public interface IColorRecognizer {
+
+	Color getColorAt(File image, int u, int v, int width, int height);
+}

src/main/java/es/yonatan/tfg/recognizer/TesseractRecognizer.java

@@ -39,7 +39,7 @@ public final class TesseractRecognizer implements IRecognizer {
 									word.getText(),
 									word.getConfidence(),
 									BoundingBox.fromRect(word.getBoundingBox()
-									), null)));
+									))));
 
 			if (iterator.hasNext())
 				units.add(Unit.NEXT_IMAGE_UNIT);