ALEXNet Optimized Cross layer Convolution Neural Network Algorithm for Efficient Diabetic Foot Ulcer Identification

Increasing variety of disease in daily fact affects the human nature, as far diabetic is most harmonic disease creates illness to affect the health. In such a way Diabetic Foot Ulcer (DFU) cause complication which is not controlled to leads diabetic treatment failures. By identifying the disease accurately is the most challenging factor because of prevailing techniques doesn’t concentrates the peripheral segmentation through image processing. This leads inaccurate to detect the disease effectively to find the risk factor. Still, the heterogeneity of diabetic skin lesions and the presence of noise in images taken with digital cameras make lesion extraction a difficult task. To resolve this problem, we presents a new method of Cross Layer Convolution Neural Network (CLCNN) is employed to improve the performance accuracy to contribute feature evaluation perfectly in segmentation level to find the ulcer. Initially preprocessing was carried out by Weiner Filter (WF) and Z-Score Monochrome Normalization (ZSMN) is applied to remove the noise to enhance the image quality. Then Color Quantum Histogram Evaluation (CQHE) is applied to find the spectral margins of wound region to increase the object entity parts findings, ALEXNet optimized Cross layer convolution neural network with ReLU activation function to increase the identification accuracy. Therefore, the experimental result provides high accuracy, sensitivity, specificity, ROC and F-score performance.