Research on false eyewitness Detection algorithm of Asian giant hornet Image based on support Vector Machine
Jiaxu Li · Journal of Image Processing Theory and Applications · 2021
Recently, Vespa mandarinia have been found frequently in Washington state, causing severe potential damage to the local ecosystem. To establish effective pest management programs, Washington State collected reports of people witnessing these wasps. However, due to the existence of a large number of wrong reports and the limited resources of government agencies, it is not possible to conduct field visits to all reports. In recognition of report images, since the image data set is extremely unbalanced, we use data enhancement, cutting and rotating the positive images to increase the number of positive images. In addition, due to the fact that the traditional neural network is easy to overfit in this kind of imbalanced data set, we use the One-Class SVM model to transform the classification problem into the outlier test problem and the experimental results show that the accuracy of our algorithm is 98%.