Ethan’s Status Report For 09/21

What did you personally accomplish this week on the project?

• Began Development Work On Several Sub-modules Of The Project
  • Completed testing with the arrays of Yolo Models to determine which one would be optimal.
  • Began work on CV Preprocessing Scripts locally so that YOLO would ONLY be run when there was actual movement.
  • Began work and have multiple completed versions of audio-to-text translators. All of which were tested with data.
• Worked on and reviewed proposal presentation
  • I created content for slides and coordinated with group members with regard to feedback for their content.

Is your progress on schedule or behind? If you are behind, what actions will be
taken to catch up to the project schedule?

I have made significant progress in my submodules by sinking a very significant amount of time into researching and implementing said components so I don’t believe I am behind. To ensure I stay ahead I will make sure to continue to be transparent with my team members about my progress and continue to improve my current designs to ensure all key requirements are met.

What deliverables do you hope to complete in the next week?

During the past week, my team spent most of our time figuring out the exact architecture of our project. Thus it was hard to do any interaction testing. However after several discussions we now how a concrete idea of what exactly we must do – whether it is implementing specific components or figuring out what approach we should take. That allows me to develop my submodules with more confidence and conduct comprehensive testing at the unit and ultimately at the integration level.

My main objective outlined have to do with getting the initial object detection model to be as accurate as possible. The major parts of this objective are:

1. Ensure there is a sizeable dataset of rooms with various levels of pixel resolution.
2. There should be easy-to-access and understandable data on how each model does with said images.
3. Next steps MUST be identified and pursued to the greatest extent possible to ensure accuracy remains high. Things like:
   1. Multiple runs of YOLO on various parts of the image zoomed in/out
   2. Additional training

I also have a secondary objective of figuring out how exactly the audio-to-text-to-image in database matching pipeline will work. This is a secondary concern as this is not the core of our product but there will likely be research into this week with a document outlining the findings being created.

Lastly, I will be working on our design review presentation to ensure it is done in time for the deadline next week.