针对被检测目标在视角变化和遮挡时较难识别的问题, 提出联合利用 Gabor特征和视角变换时共有的 LIOP特征对目标进行多角度识别的新算法。首先, 用 4个方向、 16个尺度的二维 Gabor滤波器组对输入图像进行滤波, 得到 64组含有方向信息的 Gabor特征响应图, 进而对相邻尺度和相应位置计算局部响应最大值, 得到具有尺度及平移不变的特征向量。其次, 通过几何变换算法获得不同视角下的 LIOP特征向量。然后, 为了降低时间复杂度, 通过主成分分析算法对联合特征降维。最后, 把降维后的特征向量输入支持向量机 (SVM)进行训练学习, 得到检测器模型。为了定量评估算法精度和鲁棒性, 在 Caltech-101和 UIUC car两个标准数据库进行测试, 实验结果表明, 本文在两个标准数据集上的平均识别率分别达到了 92.1%和 95.4%, 能较好检测不同尺度、不同角度的目标。

In order to solve the challenging problems of recognizing object in the angles changing and occlusion, a novel multi-angle algorithm is proposed by combining the Gabor feature and shared LIOP (Local Intensity Order Pattern) feature during the changing poses. First of all, the input image is filtered by a 2D Gabor filter in 4 directions and 16 scales to obtain 64 groups of characteristic response map. And then the scale and translation invariant feature can be derived from computing the maximum response value among the adjacent scales and position. Secondly, the geometric transformation algorithm is utilized to gain the shared LIOP feature under different perspectives. Thirdly, for reducing the time complexity, the dimension of combined features is reduced by the principal component analysis. At last, the calculated feature is trained and learned in SVM for the detecting model. Two standard test databases, Caltech 101 and UIUC car, are introduced to evaluate the accuracy and robustness of the proposed algorithm. The experimental results show that the average precision on two standard databases reach 92.1% and 95.4% high respectively, indicating the excellent performance of the proposed algorithm in recognizing object under various scales and angles.