In this project I will classify the SVHN data set to 92.77% accuracy, as well as create a number detection method. The second task is accomplished by the creation of an 11th class: background, which was has been painstakingly generated over the course of 15 hours, by hand, by myself.
DCurro/SVHN-Classification-and-Detection
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|