Senior Design Projects
ECS193 A/B Winter & Spring 2021
Developing automated methods to identify ADHD-relevant behaviors from video
| ********** | |
| Name | Meghan Miller |
| Affiliation | MIND Institute/Department of Psychiatry |
Project's details
| Project title | Developing automated methods to identify ADHD-relevant behaviors from video |
| Background | Attention-deficit/hyperactivity disorder (ADHD) is a common and impairing disorder, affecting 5-9% of children. Although it is typically diagnosed around age 7, it is thought to emerge much earlier. Our research is focused on identifying early behavioral markers of ADHD in infancy and early childhood. To do this, we prospectively follow cohorts of infants and high and low familial risk for ADHD. We examine their behavior and development between 6-36 months of age to see if we can identify any behaviors that distinguish those who develop high ADHD symptoms from those who develop typically. Additional information about the project can be found here: https://health.ucdavis.edu/mindinstitute/research/miller-lab-for-infants-with-autism-adhd/index.html |
| Description | Using a standardized coding system, our laboratory has hand-coded videos for specific ADHD-relevant behaviors in infants including grabbing items from adults, leaving their seat, and inattention/distractibility (see attached publication). Hand coding videos takes quite a lot of time and is very inefficient. We would like to develop automated methods for detecting these same behaviors from videos using computer vision. The ultimate goal is that such automated methods will allow us to identify early markers of ADHD much more quickly and easily than currently possible. |
| Deliverable | The goal is to develop initial methods for detecting one of the hand-coded behaviors, likely 'out of seat' behavior as it is the most clear and discrete of the behaviors we have coded from the videos thus far. If successful, we can compare the frequency of this behavior as detected via the computer vision algorithm to the frequency of the behavior as determined by the hand-coding that has already been completed. |
| Skill set desirable | The ideal team will have experience with computer vision techniques, as the PI does not have experience in this area. |
| Phone number | ********** |
| Client time availability | 30-60 min weekly or more |
| IP requirement | Client wishes to keep IP of the project |
| Attachment | Click here |
| Selected | No |
| Team members | N/A |
| TA | N/A |