Multiview Locally Linear Embedding for Effective Medical Image Retrieval

Abstract

Content-based medical image retrieval continues to gain attention for its potential to assist radiological image interpretation and decision making. Many approaches have been proposed to improve the performance of medical image retrieval system, among which visual features such as SIFT, LBP, and intensity histogram play a critical role. Typically, these features are concatenated into a long vector to represent medical images, and thus traditional dimension reduction techniques such as locally linear embedding (LLE), principal component analysis (PCA), or laplacian eigenmaps (LE) can be employed to reduce the “curse of dimensionality”. Though these approaches show promising performance for medical image retrieval, the feature-concatenating method ignores the fact that different features have distinct physical meanings. In this paper, we propose a new method called multiview locally linear embedding (MLLE) for medical image retrieval. Following the patch alignment framework, MLLE preserves the geometric structure of the local patch in each feature space according to the LLE criterion. To explore complementary properties among a range of features, MLLE assigns different weights to local patches from different feature spaces. Finally, MLLE employs global coordinate alignment and alternating optimization techniques to learn a smooth low-dimensional embedding from different features. To justify the effectiveness of MLLE for medical image retrieval, we compare it with conventional spectral embedding methods. We conduct experiments on a subset of the IRMA medical image data set. Evaluation results show that MLLE outperforms state-of-the-art dimension reduction methods.

Manifold-Based Visual Object Counting

“…The LLE method has been applied to multi-view and cross-modal applications [33], [34], [35], [36], where multi-modal features are exploited for image retrieval, classification or regression problems. Different from these works, however, the proposed M-VOC method focuses on modelling and deriving the correspondence from images to their density maps.…”

Large-scale image retrieval based on multi-feature and multi-kernel hashing learning

“…目前, 由于医院里每天会产生大量的医学图像, 这使得医学图像诊断的相关工作人员的工作压力 非常大. 虽然现有的 CBIR 系统在一定程度上缓解了工作人员的压力 [24,27] , 但是该系统无法应用于大 规模的医学图像的检索中, 而且其精度也有待提高. 随着基于哈希学习的图像检索技术的出现, 该问 题逐渐得到解决.…”

“…多特征的融合 解决了单特征所包含的信息量单一不足的问题; 多核的组合方式能够弥补单核学习能力上的不足且解 决了 "维度灾难" 问题. 实验结果表明本文的方法具有多特征融合 [24] 和多核学习 [25,26]…”

“…对于上述存在的问题, 本文提出一种自适应的多特征多核的哈希学习方法, 该方法先将每一种特 征映射到不同的核空间中, 然后将这些高维核空间的特征进行融合构造出一种多特征多核的哈希学习 方法. 在该方法中, 不同特征和不同核都有不同的权重, 这种方法的思想结合了多视图局部线性嵌入 方法 [24] 和多核学习方法 [25,26] , 该方法具有多特征和多核的双重优点. 本文的主要创新点在于:…”

Multi-view kernel completion

“…Learning the reconstruction weight matrix in multi-view setup has recently been considered by Shen et al (2013).…”