GTCNN

This project focues on microalgae classification. Due to a great variety of microalgae and low-qulity images, it is difficult to gain a satifying accuracy. To tackle the challenge, we propose a multi-CNNs model, named GT-CNN, to classify microalgae. Furthermore, we propose two methods to optimize the computation overhead of microalgae. Experiment results with pubilic datasets show that GT-CNN get a higher accuracy compared with the state-of-art microalgae classification methods.