The efforts to increasing the quantity and quality of rice production are obstructed by the paddy disease. This research attempted to identify the four major paddy diseases in Indonesia (leaf blast, brown spot, bacterial leaf blight, and tungro) using fractal descriptors to analyze the texture of the lesions. The lesion images were extracted manually. The descriptors of `S' component of each lesion images then used in classification process using probabilistic neural networks. This techniques achieved at least 83.00% accuracy when identifying the diseases. This method has a potential to be used as one of the feature if it combined with other features, especially when two diseases with relatively same color involved.

This research developed a mobile application for paddy plant disease identification system using fuzzy entropy and classifier probabilistic neural network (PNN) that runs on Android mobile's operating system. Paddy diseases are a major cause of yield loss and lower profit in rice production. Paddy diseases are extracted from digital paddy leaf images using fuzzy entropy and then the diseases are classified using PNN. Cross validation is used for assessing how the results of a statistical analysis will generallize to an independent data set. The experiment result shows that the accuracy of paddy diseases identification is 91.46%.

MedLeaf: Mobile Application for Medicinal Plant Identification Based on Leaf Image

This research recognizes the leaf shape using Centroid Contour Distance (CCD) as shape descriptor. CCD is an algorithm of shape representation contour-based approach which only exploits boundary information. CCD calculates the distance between the midpoint and the points on the edge corresponding to interval angle. Leaf shapes that included in this study are ellips, cordate, ovate, and lanceolate. We analyzed 200 leaf images of tropical plant. Each class consists of 50 images. The best accuracy is obtained by 96.67%. We used Probabilistic Neural Network to classify the leaf shape. Experimental results demonstrated the effectiveness of the proposed approach for shape recognition with high accuracy.

Selection of leaf features that are appropriate for identification is very important. Local Binary Patterns (LBP) is one of texture feature that is efficiency and robustness for plant identification. Meanwhile, in LBP we have to define good size of sampling point. In this research we propose fusion of LBP features, which incorporates different size of sampling point. There are two ways for fusion of LBP. First, we perform a straightforward fusion by calculating histogram of multiple LBP features separately using varying the size of sampling points and radius, then concatenate the multiple histograms together. Second, each histogram of LBP features is classified, and the feature fusion can be accomplished by classifier combination. For leaf classification, we used probabilistic neural network (PNN) to classify LBP features. The experiment performed on tropical medicinal plants and house plants. According to experimental results, the fusion of LBP features can improve accuracy in plant identification. This system is very promising to help people identify medicinal plant automatically and for conservation and utilization of medicinal plants.

Leaf Morphological Feature Extraction of Digital Image Anthocephalus Cadamba