Department Digital Imaging, Philips Research Europe, Hamburg, Hamburg, Germany.
Int J Comput Assist Radiol Surg. 2013 Jul;8(4):593-606. doi: 10.1007/s11548-013-0817-7. Epub 2013 Feb 9.
This paper proposes the discriminative generalized Hough transform (DGHT) as an efficient and reliable means for object localization in medical images. It is meant to give a deeper insight into the underlying theory and a comprehensive overview of the methodology and the scope of applications.
The DGHT combines the generalized Hough transform (GHT) with a discriminative training technique for the GHT models to obtain more efficient and robust localization results. To this end, the model points are equipped with individual weights, which are trained discriminatively with respect to a minimal localization error. Through this weighting, the models become more robust since the training focuses on common features of the target object over a set of training images. Unlike other weighting strategies, our training algorithm focuses on the error rate and allows for negative weights, which can be employed to encode rivaling structures into the model. The basic algorithm is presented here in conjunction with several extensions for fully automatic and faster processing. These include: (1) the automatic generation of models from training images and their iterative refinement, (2) the training of joint models for similar objects, and (3) a multi-level approach.
The algorithm is tested successfully for the knee in long-leg radiographs (97.6 % success rate), the vertebrae in C-arm CT (95.5 % success rate), and the femoral head in whole-body MR (100 % success rate). In addition, it is compared to Hough forests (Gall et al. in IEEE Trans Pattern Anal Mach Intell 33(11):2188-2202, 2011) for the task of knee localization (97.8 % success rate). Conclusion The DGHT has proven to be a general procedure, which can be easily applied to various tasks with high success rates.
本文提出了判别广义霍夫变换（DGHT），作为一种在医学图像中进行目标定位的高效可靠方法。本文旨在深入探讨其基本原理，全面概述其方法和应用范围。
DGHT 将广义霍夫变换（GHT）与判别式训练技术相结合，对 GHT 模型进行训练，以获得更高效、更鲁棒的定位结果。为此，模型点配备了个体权重，这些权重通过最小化定位误差进行有判别性的训练。通过这种加权，模型变得更加健壮，因为训练侧重于一组训练图像中目标对象的常见特征。与其他加权策略不同，我们的训练算法关注的是错误率，并允许使用负权重，从而可以将竞争结构编码到模型中。本文介绍了基本算法及其几个扩展，用于实现完全自动化和更快的处理。这些扩展包括：（1）从训练图像中自动生成模型，并对其进行迭代细化；（2）对相似对象的联合模型进行训练；（3）多层次方法。
该算法在长骨射线照片中的膝关节（成功率 97.6%）、C 臂 CT 中的椎骨（成功率 95.5%）和全身磁共振中的股骨头（成功率 100%）中得到了成功测试。此外，它还与霍夫森林（Gall 等人，IEEE Trans Pattern Anal Mach Intell 33(11):2188-2202, 2011）在膝关节定位任务中进行了比较（成功率 97.8%）。结论：DGHT 已被证明是一种通用的方法，可以轻松应用于各种具有高成功率的任务。
