Image recognition CNN that recognizes and tags photos of invasive species in Ontario, created at DeltaHacks 6.
The convolutional neural network or CNN that is designed and trained to recognize invasive species in Ontario was created using Python, TensorFlow and Keras. The network in its current state boasts a success rate of about 65% on test data.
The netword was trained using our own custom made data set, containing the top 8 most invasive species in ontario and some photos which do not contain an invasive species. In total, this custom data set contained around 3000 images.
The website created to hoast and integrate with the CNN was created using Flask, to facilitate the integration of the CNN and also served HTML CSS.