In the realm of healthcare, particularly in the diagnosis of Autism Spectrum Disorder (ASD), a groundbreaking study has emerged. Traditionally, diagnosing ASD has been a domain reliant on the expertise of specialized professionals, a process that is often exhaustive and not universally accessible. This has led to significant delays in diagnosis and intervention, affecting long-term […] The post Pioneering ASD Diagnosis Through AI and Retinal Imaging appeared first on Unite.AI.

  • @[email protected]
    link
    fedilink
    English
    11 year ago

    The reported average area under the receiver operating characteristic curve (AUROC) of 1.00 indicates a near-perfect ability of the models to distinguish between individuals with ASD and those with typical development.

    Highly suspicious that by looking at retinal images only, an AI model would make nearly the exact same errors as a human expert doing the assessment.