Plants exist everywhere we live, as well as places without us. Many of them carry significant information for the development of human society. The urgent situation is that many plants are at the risk of extinction. So it is very necessary to set up a database for plant protection. We believe that the first step is to teach a computer how to classify plants. Compared with other methods, such as cell and molecule biology methods, classification based on leaf image is the first choice for leaf plant classification. Sampling leaves and photoing them are low-cost and convenient. One can easily transfer the leaf image to a computer and a computer can extract features automatically in image processing techniques. Some systems employ descriptions used by botanists. But it is not easy to extract and transfer those features to a computer automatically. We have developed an efficient algorithm for leaf classification that combines high-order statistics of image features together with shape information and neural network as nonlinear classifier. The code has been tested with FLAVIA database achieving an excellent recognition rate of 92.09% (32 classes, 40 training images and the remaining images used for testing for each class, hence there are 1280 training images and 627 test images in total randomly selected and no overlap exists between the training and test images). Our approach outperforms FLAVIA algorithm and moreover it does not require any human interfered part. In FLAVIA algorithm in fact you need to mark the two terminals of the main vein of the leaf via mouse click. The distance between the two terminals is defined as the physiological length.
|File Size||24.95 kB|
|Operating System||Windows XP Windows NT Windows Windows 7 Windows 8 Windows 2000 Windows 2003 Windows 98 Windows Server 2008 Windows Vista Windows Me|