For a text image input, I need to break the text into segments using the OPENCV library
Let’s say the image has 4 lines of text, I need to write a function that breaks down and cuts the lines and creates 4 new images for each line respectively.

Then a function that receives a line/sentence that consists of several words and cuts images of the words separately.

Then a function that receives an image of a word and decomposes it into letters, for each letter its own image will be created.

The function below cutting only one word to letters

# Cutting one word to letters
def image_to_text(image_file_path):
        img = cv2.imread(image_file_path, cv2.IMREAD_COLOR)
        extracted_text = ""

        if img is not None:
            gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
            _, thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY_INV + cv2.THRESH_OTSU)
            contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

            for idx, contour in enumerate(contours):
                x, y, w, h = cv2.boundingRect(contour)
                letter_img = thresh[y:y + h, x:x + w]
                kernel = np.array([[0, -1, 0], [-1, 5, -1], [0, -1, 0]], np.float32)
                letter_img = cv2.filter2D(letter_img, -1, kernel)
                letter_img = cv2.copyMakeBorder(letter_img, 4, 4, 4, 4, cv2.BORDER_CONSTANT, value=[0, 0, 0])
                letter_img = cv2.resize(letter_img, (28, 28), interpolation=cv2.INTER_AREA)

                # Display the image
                plt.imshow(letter_img, cmap='gray')
                plt.axis('off')
                plt.show()
                letter_img = letter_img.reshape(1, 28, 28, 1).astype('float32') / 255

                prediction = loaded_model.predict(letter_img)
                # Print the predicted class and its corresponding label
                predicted_class = np.argmax(prediction)
                predicted_label = class_labels[predicted_class]
                print(f"Predicted class for the letter: {predicted_class} ({predicted_label})")
                extracted_text += predicted_label
        extracted_text = extracted_text[::-1]
        print('Extracted text:', extracted_text)
        return extracted_text

text = image_to_text(r"C:UsersstudentDesktopFinalProjectFlaskimagetest3IMG.jpg")

input:

output:

I am limited in the pictures I can upload to the forum, so I will only upload two examples of the letters B and E
Letter B from ‘Because’
Letter C from ‘Because’
The rest of the letters were extracted in the same way..

enter image description here

the question is how can i handle an input like this:
enter image description here

I’m trying for days and I’m getting frustrated.. will be glad for help