Open beefoo opened 1 month ago
This issue depends on #8
Note: Working on initial scripting but Issue #7 still depends on #8; programming with modularity in mind
Recent commit focused on what was described in today's meeting: Running the models and output their predictions into a folder. More work to come, including model metrics and formal research.
Here's a link to a working Document-- Will be adding details
Evaluate at least the models listed here:
Write a script/notebook that reads a sample set of images, and outputs annotated versions of those images with bounding boxes, labels, and scores for all objects detected.
We will need some way to visualize the results, e.g. side-by-side comparison of the results of each of the models
Create a very short write-up or "report" that summarizes the results and recommends a particular model to use.