AESPNet: Attention Enhanced Stacked Parallel Network to improve automatic Diabetic Foot Ulcer identification

XAI-FusionNet: Diabetic foot ulcer detection based on multi-scale feature fusion with explainable artificial intelligence

“…Das et al [ 39 ] developed a robust CNN-based system (AESPNet) to identify DFU ( 2023 ). To distinguish between DFU and normal skin, they arranged convolution layers in parallel and used the attention module.…”

“… Accuracy (99.06 %), F1 score (99.08 %), and Recall (98.18 %) Thotad et al [ 38 ], 2023 To build up an efficient system, named EfficientNet, three crucial parameters—width, depth, and resolution—of the CNN classifier to detect ulcers A fine-tuner unit could be added to the proposed EfficientNet to improve the training time. Precision (99 %), F1 score (98 %), Accuracy (98.97 %), and Recall (98 %) Das et al [ 39 ], 2023 An effective framework (AESPNet) to detect DFU by combining varying-sized kernel-based parallel convolution layers and a bottleneck attention module. One potential limitation of the AESPNet is the lack of explainability of the detection process.…”

“…In Ref. [ 39 ], a CNN-based AESPNet architecture was designed, consisting of 35 CNN layers and 2 Bottleneck Attention Modules. In Ref.…”