独立子空间中的场景特征增量学习方法  

Incremental Learning Towards Scene Features in Independent Subspace

作　　者：谢昭[1] 凌然[1] 吴克伟[1] 

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009

出　　处：《计算机研究与发展》2013年第11期2287-2294,共8页Journal of Computer Research and Development

基　　金：国家自然科学基金项目(60905005;60875012;61273237);高等学校博士学科点专项科研基金项目(20090111110015)

摘　　要：针对特征提取与场景描述在场景分类任务中的重要性,提出了一种独立子空间内的场景特征增量学习方法,采用基于独立子空间分析的无监督学习方法获取结构化的特征基元,基元的优化过程融入增量学习的思想框架中,以解决大样本以及动态样本下的学习难题.通过特征基元的非线性映射获取一种规则网格划分下的图像块状描述子,最后结合空间金字塔匹配模型构建层次化的场景描述,有效提高了场景图像分类的精确度.在OT场景图像集上的实验结果表明,所得特征基元能够用于构建低维高效的场景描述,通过详细讨论相关参数对优化过程以及分类性能的影响,并与多种典型模型下的实验结果进行对比,充分验证了该方法在场景分类任务中的有效性.Scene classification is not an easy task owing to the variability, ambiguity, and the wide range of illumination and scale conditions the scenes may apply. Since feature extraction and scene representation play important roles in classification tasks, this paper presents an approach for unsupervised feature learning based on independent subspace analysis. The proposed method could automatically learn structural feature bases organized in a grouped fashion from randomly sampled natural image patches in independent subspaces. Optimization process of feature bases is implemented under an incremental learning framework to cope with the learning difficulty with large or dynamic samples. Patch-based image descriptors are computed over regularly divided grids using nonlinear combination coefficients of the learned feature bases. These descriptors are then taken into the spatial ）yramid matching model, which incorporates spatial -＇orrespondence for recognizing scene categories, to layout information and global geometric ~xperiment reveals how the related parameters influence build hierarchical scene representations. objective optimization process and the final lassification performance. Compared with several typical models in classification task on OT scene ataset, the proposed method could form low-dimensional but efficient image patch descriptors and thieve high classification accuracy with stability.

关 键 词：独立子空间分析 增量学习 特征基元 空间金字塔匹配 场景分类 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]