Leveraging on the machine learning technology advances to automate plant identification and classification.
PlantCLEF 2019 was an international image-based plant identification competition organised by CIRAD, France focusing on the plant species of the Amazon rainforest and Guiana Shield, known to be the largest collection of plant and animal species in the world. The objective was to predict 10,000 different plant species given some training images. The team at NEUON managed to bag first place in the competition. The trained model was the best performing machine model in the challenge and even outperform human experts in some categories.
This video explains the overview of plant identification, its challenges and how it is done. Our team investigates the use of fine-tuned Inception-v4 and Inception-ResNet-v2 models to automate the classification of 10,000 plant species. Prior to training the networks, the training dataset was pre-processed to remove the noisy data. The team submitted three runs which achieved comparable performances to human experts on the test dataset comprising 745 observations for all the evaluation metrics. For the trained systems to generalise better, the systems were trained for multi-task classification and is able to classify plant images based on their species, with support of their genus and family labels. In particular, an ensemble of Inception-v4 and Inception-ResNet-v2 networks achieved a Top-1 accuracy of 0.316 and 0.246 for the test set identified by experts and the whole test set respectively.
We will keep you update with our latest AI technology and applications