11/9 Week Updates
The Data team received new data this week, allowing them to process it in the pipeline. The new recording was around 6 minutes. The team was able to reach 85 percent accuracy with their implemented pipeline, which is a high accuracy rate. Aside from that, the team worked on the presentation slide, preparing for the final submission.
Machine Learning Team
The ML team was able to reach 85-90% accuracy with the new data. They also uploaded a new version of the model into the git repository which will be public soon. The team is mostly tying up loose ends and preparing for the final film presentation on Tuesday.
The UI team was able to fix all the issues from last week, including changing the prediction from 1/0 to yes/no and changing the countdown button. They were able to test the hardware with the UI this week as well. The team encountered a problem where when the nodes were not connected, the response detected would always be yes, but when the nodes connected to the face, the response detected would always result in no. So this week, they will be determining the root of that problem. This week, they will also be working on the presentation slides.
- 3 toasts