Skip to content

Model Implementation Documentation

Goals of the model

  • Take in an image taken by the camera and return null, or a bounding box that has been calculated for the image

CDR

Current Progress on the model

  • Proof of concept model implemented in sauron
  • Hardcoded image file
  • detects baseball players and draws bounding boxes
  • then displays the image with bounding box overlay
  • The model is a yolo model in ONNX format
  • This model also had its ONNX Intermediate Representation (IR) changed to a lower version

Brainstorm and Plan for what to do next

Brainstorm

  • Train New Model
  • Make a jupyter notebook for this
  • Split the Model Proof of Concept into classification and visualization

Resources