改进型SIFT算法在BOW模型图像分类中的应用研究

Research on the Improved SIFT Algorithm Applied in the Image Classification Based on BOW Model

为了获得较高的图像分类正确率,针对传统SIFT算法在提取图像边缘特征及受光照影响大的缺陷,提出了一种改进型SIFT算法。实验表明,改进型SIFT算法在BOW模型中对图像的分类效果较原算法有明显的提升。

It had proposed a modified SIFT algorithm to improve previous SINF in edge feature extraction and illumination change for a higher image classification accuracy. Results showed that modified SIFT al-gorithm in BOW achieved higher classification accuracy.