Can Synthetic Images Improve CNN Performance in Wound Image Classification?

For artificial intelligence (AI) based systems to become clinically relevant, they must perform well. Machine Learning (ML) based AI systems require a large amount of labelled training data to achieve this level. In cases of a shortage of such large amounts, Generative Adversarial Networks (GAN) are a standard tool for synthesising artificial training images that can be used to augment the data set. We investigated the quality of synthetic wound images regarding two aspects: (i) improvement of wound-type classification by a Convolutional Neural Network (CNN) and (ii) how realistic such images look to clinical experts (n = 217). Concerning (i), results show a slight classification improvement. However, the connection between classification performance and the size of the artificial data set is still unclear. Regarding (ii), although the GAN could produce highly realistic images, the clinical experts took them for real in only 31% of the cases. It can be concluded that image quality may play a more significant role than data size in improving the CNN-based classification result.

Automatic Wound Type Classification with Convolutional Neural Networks.

Chronic wounds are ulcerations of the skin that fail to heal because of an underlying condition such as diabetes mellitus or venous insufficiency. The timely identification of this condition is crucial for healing. However, this identification requires expert knowledge unavailable in some care situations. Here, artificial intelligence technology may support clinicians. In this study, we explore the performance of a deep convolutional neural network to classify diabetic foot and venous leg ulcers using wound images. We trained a convolutional neural network on 863 cropped wound images. Using a hold-out test set with 80 images, the model yielded an F1-score of 0.85 on the cropped and 0.70 on the full images. This study shows promising results. However, the model must be extended in terms of wound images and wound types for application in clinical practice.

An Image Based Object Recognition System for Wound Detection and Classification of Diabetic Foot and Venous Leg Ulcers.

Venous leg ulcers and diabetic foot ulcers are the most common chronic wounds. Their prevalence has been increasing significantly over the last years, consuming scarce care resources. This study aimed to explore the performance of detection and classification algorithms for these types of wounds in images. To this end, algorithms of the YoloV5 family of pre-trained models were applied to 885 images containing at least one of the two wound types. The YoloV5m6 model provided the highest precision (0.942) and a high recall value (0.837). Its mAP_0.5:0.95 was 0.642. While the latter value is comparable to the ones reported in the literature, precision and recall were considerably higher. In conclusion, our results on good wound detection and classification may reveal a path towards (semi-) automated entry of wound information in patient records. To strengthen the trust of clinicians, we are currently incorporating a dashboard where clinicians can check the validity of the predictions against their expertise.

Visual Analytics of Gaze Data with Standard Multimedia Players.

With the increasing number of studies, where participants' eye movements are tracked while watching videos, the volume of gaze data records is growing tremendously. Unfortunately, in most cases, such data are collected in separate files in custom-made or proprietary data formats. These data are difficult to access even for experts and effectively inaccessible for non-experts. Normally expensive or custom-made software is necessary for their analysis. We address this problem by using existing multimedia container formats for distributing and archiving eye-tracking and gaze data bundled with the stimuli data. We define an exchange format that can be interpreted by standard multimedia players and can be streamed via the Internet. We convert several gaze data sets into our format, demonstrating the feasibility of our approach and allowing to visualize these data with standard multimedia players. We also introduce two VLC player add-ons, allowing for further visual analytics. We discuss the benefit of gaze data in a multimedia container and explain possible visual analytics approaches based on our implementations, converted datasets, and first user interviews.

The principal components of natural images revisited.

This paper investigates the principal components (PCs) of natural gray and color images. A horizontal and vertical typology of PCs is found which leads to the identification of groups of basis functions for steerable bandpass filters. Using this system, the contribution of spatio-chromatic structure to the total variance can be quantified for selected spatial frequencies.

The long-range saliency of edge- and corner-based salient points.

A major goal of salient-point (SP) detection is increasing computational efficiency. Therefore, methods which can detect saliency of a large region by evaluation of only a small local patch are of particular interest. This paper checks for well-known detectors whether saliency outreaches the actual SPs.

Interactive image data labeling using self-organizing maps in an augmented reality scenario.

We present an approach for the convenient labeling of image patches gathered from an unrestricted environment. The system is employed for a mobile Augmented Reality (AR) gear: while the user walks around with the head-mounted AR-gear, context-free modules for focus-of-attention permanently sample the most 'interesting' image patches. After this acquisition phase, a Self-Organizing Map (SOM) is trained on the complete set of patches, using combinations of MPEG-7 features as a data representation. The SOM allows visualization of the sampled patches and an easy manual sorting into categories. With very little effort, the user can compose a training set for a classifier, thus, unknown objects can be made known to the system. We evaluate the system for COIL-imagery and demonstrate that a user can reach satisfying categorization within few steps, even for image data sampled from walking in an office environment. (An abbreviated version of some portions of this article appeared in [Bekel, H., Heidemann, G., & Ritter, H. (2005). SOM Based Image Data Structuring in an Augmented Reality Scenario. In Proceedings of the International Joint Conference on Neural Networks, Montreal, Canada.], as part of the IJCNN 2005 conference proceedings, published under the IEEE copyright).

Focus-of-attention from local color symmetries.

In this paper, a continuous valued measure for local color symmetry is introduced. The new algorithm is an extension of the successful gray value-based symmetry map proposed by Reisfeld et al. The use of color facilitates the detection of focus points (FPs) on objects that are difficult to detect using gray-value contrast only. The detection of FPs is aimed at guiding the attention of an object recognition system; therefore, FPs have to fulfill three major requirements: stability, distinctiveness, and usability. The proposed algorithm is evaluated for these criteria and compared with the gray value-based symmetry measure and two other methods from the literature. Stability is tested against noise, object rotation, and variations of lighting. As a measure for the distinctiveness of FPs, the principal components of FP-centered windows are compared with those of windows at randomly chosen points on a large database of natural images. Finally, usability is evaluated in the context of an object recognition task.