Fine-Grained Wound Tissue Analysis Using Deep Neural Network

Abstract

Tissue assessment for chronic wounds is the basis of wound grading and selection of treatment approaches. While several image processing approaches have been proposed for automatic wound tissue analysis, there has been a shortcoming in these approaches for clinical practices. In particular, seemingly, all previous approaches have assumed only 3 tissue types in the chronic wounds, while these wounds commonly exhibit 7 distinct tissue types that presence of each one changes the treatment procedure. In this paper, for the first time, we investigate the classification of 7 wound tissue types. We work with wound professionals to build a new database of 7 types of wound tissue. We propose to use pre-trained deep neural networks for feature extraction and classification at the patch-level. We perform experiments to demonstrate that our approach outperforms other state-of-the-art. We will make our database publicly available to facilitate research in wound assessment.

A superpixel-driven deep learning approach for the analysis of dermatological wounds

“…While the segmentation of wounded tissues from photographic images of dermatological ulcers has been discussed from distinct perspectives [3,8,24], most of the current approaches perform a threestage pipeline for the labeling of wounds towards specific beacons and markers [10,11]. Such a pipeline is composed of (i) region segmentation, which aims to remove image noise and delimit region boundaries, (ii) feature extraction, which represents (parts of) an image in a multidimensional space, and (iii) data classification, which assigns a label to each image representation.…”

“…DL models have recently surpassed human performance in image classification from basic to complex tasks [17,25]. A variation of DL models is using a CNN method only for feature extraction, as in the proposal of Nejati et al [24]. Their approach uniformly divides an image into patches that are fed to five convolutional layers.…”

“…A frequent scenario in wound analysis is the uneven distribution of tissue patterns [32,33]. For instance, Nejati et al [24] evaluated a dataset of 350 wound images divided into labeled patches, where most of the labels were related to only three of seven possible classes. The authors addressed such imbalance by using an approximation of the NP-hard problem that divides the instances into training and test sets.…”

“…[23] proposed a transfer-learning DL model for diabetic foot segmentation with 92.5% accuracy, while Nejati et al [24] combined AlexNet and SVM for wound classification at an 86.40% hit ratio. Analogously, Zahia et al…”

Wound image segmentation using clustering based algorithms

“…• image classification • image segmentation Paper "Fine-grained wound tissue analysis using deep neural network" (Nejati, H., et al, 2018) describes development of the algorithm for 7 types of chronic wounds recognition. It is based on deep neural networks (DNN).…”