Automatic and accurate estimation of condition severity is essential for foods stability, illness management, and yield reduction prediction. The deep learning process avoids labor-intensive feature engineering and threshold-based mostly graphic segmentation.Their activation can for this reason be computed by using a matrix multiplication followed