ABSTRACTVaricose ulcers occur due to improper functioning of venous valves in legs. These ulcers are the severe external signs of vein-related problems, such as chronic pain, leg swelling and leg heaviness. The aim of the proposed methodology is to increase better diagnosis and treatment results by computer-assisted tissue classification (granulation, slough, necrotic and epithelial) of varicose ulcer using the Multidimensional Convolutional Neural Network as a deep learning architecture. This work consists of (i) preprocessing to remove the flash light reflection from the RGB wound images, (ii) active contour segmentation to segment the wounded areas from the skin and (iii) Multidimensional convolutional neural network for which the segmented images and their corresponding ground truth images are given as input. After CNN training, the fully connected layer gives the output as segmented images which include different types of tissues which are to be predicted. The Multidimensional Convolutional Neural Network structure of the layer can be modified by defining the layer structure using Matlab functions to get more accurate results for tissue classification. The proposed approach is evaluated using metrics with efficient performance rates of average accuracy (99.55%), specificity (98.06%) and sensitivity (95.66%). Experiments conducted on varicose ulcer wound image aim to improve healing status and skin health conditions based on the texture of the tissue.