Autism spectrum disorder (ASD) is a neurological illness characterized by deficits in cognition, physical activities, and social skills. There is no specific medication to treat this illness; only early intervention can improve brain functionality. Since there is no medical test to identify ASD, a diagnosis might be challenging. In order to determine a diagnosis, doctors consider the child’s behavior and developmental history. The human face can be used as a biomarker as it is one of the potential reflections of the brain and thus can be used as a simple and handy tool for early diagnosis. This study uses several deep convolutional neural network (CNN)-based transfer learning approaches to detect autistic children using the facial image. An empirical study is conducted to select the best optimizer and set of hyperparameters to achieve better prediction accuracy using the CNN model. After training and validating with the optimized setting, the modified Xception model demonstrates the best performance by achieving an accuracy of 95% on the test set, whereas the VGG19, ResNet50V2, MobileNetV2, and EfficientNetB0 achieved 86.5%, 94%, 92%, and 85.8%, accuracy, respectively. Our preliminary computational results demonstrate that our transfer learning approaches outperformed existing methods. Our modified model can be employed to assist doctors and practitioners in validating their initial screening to detect children with ASD disease.

The research describes an effective deep learning-based, data-centric approach for diagnosing autism spectrum disorder from facial images. To classify ASD and non-ASD subjects, this method requires training a convolutional neural network using the facial image dataset. As a part of the data-centric approach, this research applies pre-processing and synthesizing of the training dataset. The trained model is subsequently evaluated on an independent test set in order to assess the performance matrices of various data-centric approaches. The results reveal that the proposed method that simultaneously applies the pre-processing and augmentation approach on the training dataset outperforms the recent works, achieving excellent 98.9% prediction accuracy, sensitivity, and specificity while having 99.9% AUC. This work enhances the clarity and comprehensibility of the algorithm by integrating explainable AI techniques, providing clinicians with valuable and interpretable insights into the decision-making process of the ASD diagnosis model.

Autism Spectrum Disorder (ASD) is a developmental disorder that impacts social communication and conduct. ASD lacks standard treatment protocols or medication, thus early identification and proper intervention are the most effective procedures to treat this disorder. Artificial intelligence could be a very effective tool to be used in ASD diagnosis as this is free from human bias. This research examines the effect of face alignment for the early diagnosis of Autism Spectrum Disorder (ASD) using facial images with the possibility that face alignment can improve the prediction accuracy of deep learning algorithms. This work uses the SOTA deep learning-based face alignment algorithm MTCNN to preprocess the raw data. In addition, the impacts of facial alignment on ASD diagnosis using facial images are investigated using state-of-the-art CNN backbones such as ResNet50, Xception, and MobileNet. ResNet50V2 achieves the maximum prediction accuracy of 93.97% and AUC of 96.33% with the alignment of training samples, which is a substantial improvement over previous research. This research paves the way for a data-centric approach that can be applied to medical datasets in order to improve the efficacy of deep neural network algorithms used to develop smart medical devices for the benefit of mankind. ABSTRAK: Gangguan Spektrum Autisme (ASD) adalah gangguan perkembangan yang memberi kesan kepada komunikasi dan tingkah laku sosial. Kelemahan dalam rawatan ASD adalah ianya tidak mempunyai protokol rawatan standard atau ubat. Oleh itu pengenalan awal dan campur tangan betul merupakan prosedur paling berkesan bagi merawat gangguan ini. Kecerdasan buatan boleh menjadi alat berkesan bagi diagnosis ASD kerana bebas campur tangan manusia. Penyelidikan ini mengkaji kesan penjajaran muka bagi diagnosis awal ASD menggunakan imej muka dengan kebarangkalian penjajaran muka dapat meningkatkan ketepatan ramalan algoritma pembelajaran mendalam. Kajian ini menggunakan algoritma penjajaran muka MTCNN berasaskan pembelajaran mendalam SOTA bagi pra-proses data mentah. Selain itu, kesan penjajaran muka pada diagnosis ASD menggunakan imej muka disiasat menggunakan CNN terkini seperti ResNet50, Xception dan MobileNet. ResNet50V2 mencapai ketepatan ramalan maksimum sebanyak 93.97% dan AUC 96.33% dengan sampel penjajaran latihan, yang merupakan peningkatan ketara berbanding penyelidikan terdahulu. Kajian ini membuka jalan bagi pendekatan data berpusat yang boleh digunakan pada set data perubatan bagi meningkatkan keberkesanan algoritma rangkaian saraf mendalam dan membangunkan peranti perubatan pintar bermanfaat untuk manusia.

The early diagnosis of autism spectrum disorder (ASD) encounters challenges stemming from domain variations in facial image datasets. This study investigates the potential of active learning, particularly uncertainty-based sampling, for domain adaptation in early ASD diagnosis. Our focus is on improving model performance across diverse data sources. Utilizing the Kaggle ASD and YTUIA datasets, we meticulously analyze domain variations and assess transfer learning and active learning methodologies. Two state-of-the-art convolutional neural networks, Xception and ResNet50V2, pretrained on distinct datasets, demonstrate noteworthy accuracies of 95% on Kaggle ASD and 96% on YTUIA, respectively. However, combining datasets results in a modest decline in average accuracy, underscoring the necessity for effective domain adaptation techniques. We employ uncertainty-based active learning to address this, which significantly mitigates the accuracy drop. Xception and ResNet50V2 achieve 80% and 79% accuracy when pretrained on Kaggle ASD and applying active learning on YTUIA, respectively. Our findings highlight the efficacy of uncertainty-based active learning for domain adaptation, showcasing its potential to enhance accuracy and reduce annotation needs in early ASD diagnosis. This study contributes to the growing body of literature on ASD diagnosis methodologies. Future research should delve deeper into refining active learning strategies, ultimately paving the way for more robust and efficient ASD detection tools across diverse datasets.

This paper aims to investigate the effectiveness of different modalities of facial images for diagnosing Autism Spectrum Disorder (ASD) using deep learning-based neural networks. The motivation behind this study is the potential of advanced technologies to aid in accurately diagnosing ASD. The research revolves around the need to explore the performance of deep learning models on different modalities of facial images and to identify the challenges and potential solutions associated with each modality. The methodology involves training and testing the models on the respective datasets and analysing their accuracy and performance. ResNet50V2 achieved a 100% accuracy on the 2D test dataset, while Xception achieved an accuracy of 93.75% on the 3D test set. The detection accuracy suggests that neural networks-based deep learning methods have the potential to diagnose ASD using facial images accurately. However, the models perform better on 2D data, highlighting the need for additional training on larger 3D datasets to improve accuracy on 3D images. The study contributes to the field by providing insights into the performance of different modalities of facial images, emphasizing the need for robust datasets, and suggesting future research directions to enhance the accuracy and efficiency of ASD diagnosis using deep learning techniques.

Background/Objectives: Artificial intelligence (AI) is revolutionising healthcare for people with disabilities, including those with autism spectrum disorder (ASD), in the era of advanced technology. This work explicitly addresses the challenges posed by inconsistent data from various sources by developing and evaluating a robust deep ensemble learning system for the accurate and reliable classification of autism spectrum disorder (ASD) based on facial images. Methods: We created a system that learns from two publicly accessible datasets of ASD images (Kaggle and YTUIA), each with unique demographics and image characteristics. Utilising a weighted ensemble strategy (FPPR), our innovative ASD-UANet ensemble combines the Xception and ResNet50V2 models to maximise model contributions. This methodology underwent extensive testing on a range of groups stratified by age and gender, including a critical assessment of an unseen, real-time dataset (UIFID) to determine how well it generalised to new domains. Results: The performance of the ASD-UANet ensemble was consistently better. It significantly outperformed individual transfer learning models (e.g., Xception alone on T1+T2 yielded an accuracy of 83%), achieving an impressive 96.0% accuracy and an AUC of 0.990 on the combined-domain dataset (T1+T2). Notably, the ASD-UANet ensemble demonstrated strong generalisation on the unseen real-time dataset (T3), achieving 90.6% accuracy and an AUC of 0.930. This demonstrates how well it generalises to new data distributions. Conclusions: Our findings demonstrate significant potential for widespread, equitable, and clinically beneficial ASD screening using this promising, reasonably priced, and non-invasive method. This study establishes the foundation for more precise diagnoses and greater inclusion for people with autism spectrum disorder (ASD) by integrating methods for diverse data and combining deep learning models.