The data set the current model is using is pretty bad... the images are extremely low res making it nearly impossible even for a human to classify the character in the shape. The shape detection works perfectly though. As of commit bd97ace the (old) dataset used can be found here.
Goal/Purpose
The goal is to obtain or create a dataset that can reflect competition conditions.
User Story
As a developer, I want "good" data, so that I can train my vision system appropriately.
Description
Overview
The data set the current model is using is pretty bad... the images are extremely low res making it nearly impossible even for a human to classify the character in the shape. The shape detection works perfectly though. As of commit bd97ace the (old) dataset used can be found here.
Goal/Purpose
The goal is to obtain or create a dataset that can reflect competition conditions.
User Story
As a developer, I want "good" data, so that I can train my vision system appropriately.
Related Documents/Links
Acceptance Criteria