Towards population-independent, multi-disease detection in fundus photographs.

Independent validation studies of automatic diabetic retinopathy screening systems have recently shown a drop of screening performance on external data. Beyond diabetic retinopathy, this study investigates the generalizability of deep learning (DL) algorithms for screening various ocular anomalies in fundus photographs, across heterogeneous populations and imaging protocols. The following datasets are considered: OPHDIAT (France, diabetic population), OphtaMaine (France, general population), RIADD (India, general population) and ODIR (China, general population). Two multi-disease DL algorithms were developed: a Single-Dataset (SD) network, trained on the largest dataset (OPHDIAT), and a Multiple-Dataset (MD) network, trained on multiple datasets simultaneously. To assess their generalizability, both algorithms were evaluated whenever training and test data originate from overlapping datasets or from disjoint datasets. The SD network achieved a mean per-disease area under the receiver operating characteristic curve (mAUC) of 0.9571 on OPHDIAT. However, it generalized poorly to the other three datasets (mAUC < 0.9). When all four datasets were involved in training, the MD network significantly outperformed the SD network (p = 0.0058), indicating improved generality. However, in leave-one-dataset-out experiments, performance of the MD network was significantly lower on populations unseen during training than on populations involved in training (p < 0.0001), indicating imperfect generalizability.

Computer-Aided Screening of Autism Spectrum Disorder: Eye-Tracking Study Using Data Visualization and Deep Learning.

BACKGROUND: The early diagnosis of autism spectrum disorder (ASD) is highly desirable but remains a challenging task, which requires a set of cognitive tests and hours of clinical examinations. In addition, variations of such symptoms exist, which can make the identification of ASD even more difficult. Although diagnosis tests are largely developed by experts, they are still subject to human bias. In this respect, computer-assisted technologies can play a key role in supporting the screening process. OBJECTIVE: This paper follows on the path of using eye tracking as an integrated part of screening assessment in ASD based on the characteristic elements of the eye gaze. This study adds to the mounting efforts in using eye tracking technology to support the process of ASD screening. METHODS: The proposed approach basically aims to integrate eye tracking with visualization and machine learning. A group of 59 school-aged participants took part in the study. The participants were invited to watch a set of age-appropriate photographs and videos related to social cognition. Initially, eye-tracking scanpaths were transformed into a visual representation as a set of images. Subsequently, a convolutional neural network was trained to perform the image classification task. RESULTS: The experimental results demonstrated that the visual representation could simplify the diagnostic task and also attained high accuracy. Specifically, the convolutional neural network model could achieve a promising classification accuracy. This largely suggests that visualizations could successfully encode the information of gaze motion and its underlying dynamics. Further, we explored possible correlations between the autism severity and the dynamics of eye movement based on the maximal information coefficient. The findings primarily show that the combination of eye tracking, visualization, and machine learning have strong potential in developing an objective tool to assist in the screening of ASD. CONCLUSIONS: Broadly speaking, the approach we propose could be transferable to screening for other disorders, particularly neurodevelopmental disorders.

Variational Autoencoder for Image-Based Augmentation of Eye-Tracking Data.

Over the past decade, deep learning has achieved unprecedented successes in a diversity of application domains, given large-scale datasets. However, particular domains, such as healthcare, inherently suffer from data paucity and imbalance. Moreover, datasets could be largely inaccessible due to privacy concerns, or lack of data-sharing incentives. Such challenges have attached significance to the application of generative modeling and data augmentation in that domain. In this context, this study explores a machine learning-based approach for generating synthetic eye-tracking data. We explore a novel application of variational autoencoders (VAEs) in this regard. More specifically, a VAE model is trained to generate an image-based representation of the eye-tracking output, so-called scanpaths. Overall, our results validate that the VAE model could generate a plausible output from a limited dataset. Finally, it is empirically demonstrated that such approach could be employed as a mechanism for data augmentation to improve the performance in classification tasks.

Learning Clusters in Autism Spectrum Disorder: Image-Based Clustering of Eye-Tracking Scanpaths with Deep Autoencoder.

Autism spectrum disorder (ASD) is a lifelong condition characterized by social and communication impairments. This study attempts to apply unsupervised Machine Learning to discover clusters in ASD. The key idea is to learn clusters based on the visual representation of eye-tracking scanpaths. The clustering model was trained using compressed representations learned by a deep autoencoder. Our experimental results demonstrate a promising tendency of clustering structure. Further, the clusters are explored to provide interesting insights into the characteristics of the gaze behavior involved in autism.