Leprosy is a continuing community-health issue and chronic wounds with peripheral nerve loss are a challenging issue in the modern global care. Upon recurrence, the ulcers are characterized by slow healing patterns and tend to superin-fection; hence, difficult to manage and particularly, in resource- poor countries. It has been argued that traditional methods of diagnostic processes which involve manual inspection methods are inaccurate and thus advancement of other methods which involve the use of technology. In this case, we shall come up with a powerful framework that will employ the latest picture processing innovations to identify and categorize chronic wounds at different levels and severity. The pipeline embraces the thresholding and morphological feature that is applied to accentuate the wound at- tributes, and further segmentation through UNET++ architecture to cut out the affected area with the supply of the background. This is due to transfer learning together with VGG-19 and Inception model to produce a hybrid model to correctly classify the images thus defining the treatment regimen. Accuracy, F1 score, precision, recall, and an error rate accommodate the assessment of the models; their values are given in the form of a bar chart. There is also a web application that has been developed on the basis of a patient-focus to simplify the process of image submission, gathering of personal data, and the ability to present the results in terms of wound-type and -severity. The end goal or aim of this automated framework would eventually be to make the early detection and treatment of chronic wounds a possibility and in addition to that, to improve the outcomes of a patient and at the same time positively impact and enhance the efficiency of care.
BT - 2025 International Conference on Next Generation Computing Systems (ICNGCS) DO - 10.1109/icngcs64900.2025.11183359 LA - ENG M3 - Article N2 -Leprosy is a continuing community-health issue and chronic wounds with peripheral nerve loss are a challenging issue in the modern global care. Upon recurrence, the ulcers are characterized by slow healing patterns and tend to superin-fection; hence, difficult to manage and particularly, in resource- poor countries. It has been argued that traditional methods of diagnostic processes which involve manual inspection methods are inaccurate and thus advancement of other methods which involve the use of technology. In this case, we shall come up with a powerful framework that will employ the latest picture processing innovations to identify and categorize chronic wounds at different levels and severity. The pipeline embraces the thresholding and morphological feature that is applied to accentuate the wound at- tributes, and further segmentation through UNET++ architecture to cut out the affected area with the supply of the background. This is due to transfer learning together with VGG-19 and Inception model to produce a hybrid model to correctly classify the images thus defining the treatment regimen. Accuracy, F1 score, precision, recall, and an error rate accommodate the assessment of the models; their values are given in the form of a bar chart. There is also a web application that has been developed on the basis of a patient-focus to simplify the process of image submission, gathering of personal data, and the ability to present the results in terms of wound-type and -severity. The end goal or aim of this automated framework would eventually be to make the early detection and treatment of chronic wounds a possibility and in addition to that, to improve the outcomes of a patient and at the same time positively impact and enhance the efficiency of care.
PB - IEEE PY - 2025 SP - 1 EP - 8 T2 - 2025 International Conference on Next Generation Computing Systems (ICNGCS) TI - Healthcare for Chronic Wound Management In Leprosy Patients ER -