Orange Fruit Disease Detection Using Deep Convolutional Neural Networks
The paper investigates different deep learning techniques and gives a model for orange disease detection based on deep convolutional neural networks. The model used in this study uses a visual geometry group named convolutional neural network which achieved top performance in ILSVRC 2014 competition. Fruit disease detection has been a tough task to achieve using computers. Today machine techniques coupled with very high speed computing hardware devices have allowed the use of algorithms that automatically detect objects in an image. Utilizing these new advancement of neural networks a relatively good model for classifying images into fresh and diseased is built. The dataset is acquired from Kaggle.com which contains images for three different fruits but the images for oranges have been used in this experiment to make the model specific. Many different species of fruit will require the model to recognise the type of fruits increasing the computation cost. Instead making separate models allows the model to become specific and increase accuracy in one domain, this will generally allow better results.
Keywords: deep convolutional neural networks,Kaggle, visual geometry group