Human and computer attention in assessing genetic conditions
Duong, D.; Johny, A. R.; Ledgister Hanchard, S.; Fortney, C.; Hellmann, F.; Hu, P.; Moosa, S.; Patel, T.; Persky, S.; Sumer, O.; Tekendo-Ngongang, C.; Hsieh, T.-C.; Waikel, R. L.; Andre, E.; Krawitz, P.; Solomon, B. D.
Show abstract
Deep learning (DL) and other types of artificial intelligence (AI) are increasingly used in many biomedical areas, including genetics. One frequent use in medical genetics involves evaluating images of people with potential genetic conditions to help with diagnosis. A central question involves better understanding how AI classifiers assess images compared to humans. To explore this, we performed eye-tracking analyses of geneticist clinicians and non-clinicians. We compared results to DL-based saliency maps. We found that human visual attention when assessing images differs greatly from the parts of images weighted by the DL model. Further, individuals tend to have a specific pattern of image inspection, and clinicians demonstrate different visual attention patterns than non-clinicians.
Matching journals
The top 9 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Advancing Genotype-Phenotype Analysis through 3D Facial Morphometry: Insights from Cri-du-Chat Syndrome 92%
- EyeG2P: an automated variant filtering approach improves efficiency of diagnostic genomic testing for inherited ophthalmic disorders 90%
- How do clinician and parent reported data differ? An analysis of similarity and difference in the datasets from a cross-syndrome genetics cohort study(GenROC) 90%
Similar papers in this journal
- Early Diagnosis of Vascular Ehlers-Danlos Syndrome Through AI-Powered Facial Analysis: Results from the Montalcino Aortic Consortium 93%
- Genetic Diagnosis of Facioscapulohumeral Muscular Dystrophy Type 1 Using Rare Variant Linkage Analysis and Long Read Genome Sequencing 88%
- Evaluation of DNA-poli: study protocol of a randomised controlled trial to assess a digital platform for family cascade genetic testing and predictive genetic counselling 86%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.