In this paper, we propose a novel method for plant leaves recognition by incorporating an unsupervised convolutional auto-encoder (CAE) and Siamese neural network in a unified framework by considering Siamese as an alternative to the conventional loss of CAE. Rather than the conventional exploitation of CAE and Siamese, in our case we have proposed to extend CAE for a novel supervised scenario by considering it as one-class learning classifier. For each class, CAE is trained to reconstruct its positive and negative examples and Siamese is trained to distinguish the similarity and the dissimilarity of the obtained examples. On the contrary and asymmetric to the related hierarchical classification schemes which require pre-knowledge on the dataset being recognized, we propose a hierarchical classification scheme that doesn’t require such a pre-knowledge and can be employed by non-experts automatically. We cluster the dataset to assemble similar classes together. A test image is first assigned to the nearest cluster, then matched to one class from the classes that fall under the determined cluster using our novel one-class learning classifier. The proposed method has been evaluated on the ImageCLEF2012 dataset. Experimental results have proved the superiority of our method compared to several state-of-the art methods.

Because they are exploited in many fields such as medicine, agriculture, chemistry and others, plants are of fundamental importance to life on earth. Before it can be used, a plant need to firstly be identified and categorized. However, a manual identification task requires time, and it is not an easy task to do. This is because some plants look visually similar to the human eye, whereas some others may be unknown to it. Therefore, there has been an increasing interest in developing a system that automatically fulfils such tasks fast and accurate. In this paper, we propose an automatic plant classification system based on a parallel combination technique of multiple classifiers. We have considered using three widely known classifiers namely Naïve Bayes (NB), K-Nearest Neighbour (KNN) and Support Vector Machine (SVM). Our system has been evaluated using the well-known Flavia dataset. It has shown a better performance than those obtained using only one classifier.

No abstract