Add diagonal corner lines for better visibility

[kanawanaikoto]

Modify some function parameters to achieve this.
Original code appears to be invalid redundant code.

corner_length = min(20, w // 4, h // 4)

cv2.line(frame, (x1, y1), (x1 + corner_length, y1), color, thickness)
cv2.line(frame, (x1, y1), (x1, y1 + corner_length), color, thickness)

cv2.line(frame, (x1 + w, y1), (x1 + w - corner_length, y1), color, thickness)
cv2.line(frame, (x1 + w, y1), (x1 + w, y1 + corner_length), color, thickness)

cv2.line(frame, (x1, y1 + h), (x1 + corner_length, y1 + h), color, thickness)
cv2.line(frame, (x1, y1 + h), (x1, y1 + h - corner_length), color, thickness)

cv2.line(frame, (x1 + w, y1 + h), (x1 + w - corner_length, y1 + h), color, thickness)
cv2.line(frame, (x1 + w, y1 + h), (x1 + w, y1 + h - corner_length), color, thickness)

The deletion of the code above does not affect the realization of the original effect
Before

After

I have taken note of your comment and made some modifications.

[hassanrrraza]

Thank you for the contribution!
However, this version introduces arbitrary +2 and -2 offsets, which may cause the corner lines to look misaligned, especially on smaller bounding boxes. It also adds unnecessary complexity to otherwise clean and precise drawing logic.

That said, this approach might be useful in specific cases, such as:
-Creating visual separation from the bounding box edges
-Applying anti-aliasing or avoiding overlap with other UI elements
-Performing pixel-level tuning for aesthetic reasons

Unless we have a strong visual requirement for such tweaks, I recommend keeping the original version for consistency and clarity.

[kanawanaikoto]

Thank you for your response. I understand your intention, but in this case, the following code seems to be redundant.

corner_length = min(20, w // 4, h // 4)

cv2.line(frame, (x1, y1), (x1 + corner_length, y1), color, thickness)
cv2.line(frame, (x1, y1), (x1, y1 + corner_length), color, thickness)

cv2.line(frame, (x1 + w, y1), (x1 + w - corner_length, y1), color, thickness)
cv2.line(frame, (x1 + w, y1), (x1 + w, y1 + corner_length), color, thickness)

cv2.line(frame, (x1, y1 + h), (x1 + corner_length, y1 + h), color, thickness)
cv2.line(frame, (x1, y1 + h), (x1, y1 + h - corner_length), color, thickness)

cv2.line(frame, (x1 + w, y1 + h), (x1 + w - corner_length, y1 + h), color, thickness)
cv2.line(frame, (x1 + w, y1 + h), (x1 + w, y1 + h - corner_length), color, thickness)

When drawing the detection box, you used thickness=1, and when drawing the lines, you still used thickness=1. Since these two parts overlap, the final line width remains thickness=1, making the corner lines appear unnecessary.
Before Delete

After deletion

At least on my device, it doesn't seem to make a difference.
Can this code be deleted, or is there a specific purpose for it?