Flower or Crops CNN classification

Overview

This repository contains the training/evaluation scripts for a Convolutional Neural Network made for detecting flowers in the middle of a plantation. The goal is to detect ipomoea grandifolia, one of the weeds common in sugarcane plantations.

CNN Architectures

The idea is to reuse popular Imagenet architectures, fine-tuning them with a custom set of flower images (signal) and miscellaneous plantation images (background).