基于Bag of Words的干果图像分类研究

Research on Digital Dried Fruit Image Classification Based on Bag-of-Words Model

针对干果图像信息量大、分类精度低和耗时多的特点,提出利用Bag of Words模型提取图片的代表特征,并采用朴素贝叶斯分类器指导特征矩阵分类。结果表明,图像分类精度能达到80%,分类处理时间约为2 s。通过增加学习样本来进一步提高分类精度,将Bag of Words应用于干果图像识别和分类是可行的。

According to the characteristics of digital dried fruit image classification which have lots of information,weaken classification accuracy and more time-consuming,it is put forward to extract image representation using the Bag-of-Words model and to classify the feature matrix with Nave Bayes Classifier. The results showed that the accuracy was over 80%,the treatment time was 2 seconds. By increasing the learning samples to further improve the classification accuracy,the Bag of Words applied to the dried fruit image recognition and classification is feasible.