This is a repository for the pattern recognition class! The name of this group is called Machines Can "C" and contains the following group members:
- Keaton Brewster
- Christine Case
- Roy Salinas
The scope of this study will be encapsulated by the training and test data from Viz-Wiz, the machine learning methods gleaned in this class, and our own novel problem solving skills. The data from Viz-Wiz contains real data from the visually impaired who required assistance with visual recognition with everyday objects. The data was collected through an application that provided live assistance in classifying an object with virtual human interaction that allowed for easein gathering the data.
The scope of this project is to correctly classify the images based on the following features
- Size
- Color
- Shape
- Brightness
Questions such as the following are some of the questions we are looking to be able to answer:
What color is this?
What shape is this?
Is this bright?
What does this say?
Of course, these are fundamental questions and we will expand accordingly and update the repository as we make progress.
The input and output of this program will consist of an individual uploading an image to a webserver, have the user ask a specific question about the picture, and have an answer prompted on the webserver.