Shirley’s Status Report: Week 11

Progress

1. Trained primary yolov3 model supporting one class only (AC). Model being able to recognize AC in camera captured frame.
2. Tested image stitching before object detection. Turned out that object detection fails for panorama, suspecting the reason being the warping edges and irregular input image size. Gave up using panorama for now and focus on developing multi-image transferring & processing pipeline.
3. Collected image data for the rest of the 4 classes and start labeling.

Schedule

1. Train labeled image dataset for all classes.
2. Start establishing separate Object Detection EC2 server.
3. Start working on data post-processing for recognizing device location.

Deliverables Next Week

1. Finish training a model for all devices and deploy the weights on local server
2. Crudely calculate device location according to our discussion in week.