Question on bbox_size Calculation in _detect Method

In [src/data/VSRiW_extract_visual_features.py](https://github.com/mpc001/Visual_Speech_Recognition_for_Multiple_Languages/blob/5e1405db0ae816509fb312f9a578724c2e0de0c7/pipelines/detectors/mediapipe/detector.py#L39-L42), the `_detect` method calculates bbox_size as `(bbox[2] - bbox[0]) + (bbox[3] - bbox[1])`, where bbox is (xmin, ymin, width, height). This simplifies to `(width - xmin) + (height - ymin)`, which doesn't reflect the actual bounding box size (may even be negative).
Logically, bbox_size should be width + height (i.e., `bbox[2] + bbox[3]`) or `(xmax - xmin) + (ymax - ymin)` (with `xmax = xmin + width, ymax = ymin + height`).
Is this a mistake, or intentional? Thanks for clarifying.

https://github.com/Bose/RAVEN/blob/11dca128d80fcf67d8547bc2e03c31a38a7f2e84/src/data/VSRiW_extract_visual_features.py#L80-L83

(Edited for formatting)