Summary of Ensemble Modeling Of Multiple Physical Indicators to Dynamically Phenotype Autism Spectrum Disorder, by Marie Huynh (1) et al.
Ensemble Modeling of Multiple Physical Indicators to Dynamically Phenotype Autism Spectrum Disorder
by Marie Huynh, Aaron Kline, Saimourya Surabhi, Kaitlyn Dunlap, Onur Cezmi Mutlu, Mohammadmahdi Honarmand, Parnian Azizian, Peter Washington, Dennis P. Wall
First submitted to arxiv on: 23 Aug 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper presents a novel dataset and computer vision models for detecting autism spectrum disorder (ASD) through analyzing naturalistic home videos captured via the GuessWhat mobile application. The dataset consists of over 3,000 structured videos from 382 children, including those with and without ASD. This unique collection provides a robust foundation for training models to identify ASD-related phenotypic markers, such as variations in emotional expression, eye contact, and head movements. The authors develop a protocol for curating high-quality videos and train individual LSTM-based models using various input features, achieving test AUCs ranging from 67% to 86%. To improve diagnostic accuracy, they apply late fusion techniques to create ensemble models, boosting the overall AUC to 90%. This approach also yields more equitable results across different genders and age groups. The methodology has significant potential in reducing reliance on subjective assessments, making early identification more accessible and equitable. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper helps doctors and researchers detect autism earlier by using special computer programs that analyze videos of children playing with their parents. They used over 3,000 videos from the GuessWhat app to train these models. The program looks at things like how children express emotions, make eye contact, and move their heads. The results show that this method is quite good at detecting autism, especially when combining different techniques. This can help doctors diagnose autism more accurately and earlier, which is important because early treatment makes a big difference. |
Keywords
* Artificial intelligence * Auc * Boosting * Lstm
