Identify and Apply Text Color

[mschmermer]

Description: Translated text sometimes appears in black on dark backgrounds, making it hard to read.

Problem: When the background is dark, the translated text is placed in black, causing readability issues.

Solution:

Analyze the background color of the text area. Adjust the text color for sufficient contrast. Use light text on dark backgrounds and dark text on light backgrounds.

[mschmermer]

i found a better solution.

add a white border around the text by move the text with white color around the default text by:

def draw_text(image: np.ndarray, blk_list: List[TextBlock], font_pth: str, init_font_size, colour: str = "#000"):
    image = cv2_to_pil(image)
    draw = ImageDraw.Draw(image)

    font = ImageFont.truetype(font_pth, size=init_font_size)

    for blk in blk_list:
        x1, y1, width, height = blk.xywh
        tbbox_top_left = (x1, y1)

        translation = blk.translation
        if not translation or len(translation) == 1:
            continue

        translation, font_size = pil_word_wrap(image, tbbox_top_left, font_pth, init_font_size, translation, width, height, align=blk.alignment, spacing=blk.line_spacing)
        font = font.font_variant(size=font_size)

        offsets = [(dx, dy) for dx in (-2, -1, 0, 1, 2) for dy in (-2, -1, 0, 1, 2) if dx != 0 or dy != 0]
        for dx, dy in offsets:
            draw.multiline_text((tbbox_top_left[0] + dx, tbbox_top_left[1] + dy), translation, font=font, fill="#FFF", align=blk.alignment, spacing=1)

        draw.multiline_text(tbbox_top_left, translation, colour, font, align=blk.alignment, spacing=1)
    image = pil_to_cv2(image)
    return image

[mschmermer]

[ayoosh007]

You should create a pull request

[mschmermer]

I will do that if i have time. Feature is ready and tested so far ...

[ogkalu2]

@mschmermer Thanks, I've implemented your fix. I've tested it and i agree it works better than trying to detect the font color so i'll close this.