A study explored advanced convolutional neural networks (CNNs) to improve post-surgical wound care in anorectal diseases. The researchers assessed three architectures—MobileNet, ResNet50, and Inception-v4—focusing on wound characteristics such as size and severity. Results revealed that Inception-v4 surpassed its counterparts in accuracy, precision, and recall, demonstrating significant potential for clinical application. Additionally, integrating Grad-CAM technology enhanced interpretability, providing insights into the decision-making processes of the models, which contributes to optimized patient care strategies.
Comparative Study by Zhang Q, Chen Z, Hu S and Bao X in Ann Ital Chir