naa a22 4500 510807992 CHVBK 20180411083427.0 cr unu---uuuuu 180411e20130701xx s 000 0 eng 10.1007/s11548-013-0817-7 doi (NATIONALLICENCE)springer-10.1007/s11548-013-0817-7 Discriminative generalized Hough transform for object localization in medical images [Elektronische Daten] [Heike Ruppertshofen, Cristian Lorenz, Georg Rose, Hauke Schramm] PurposeThis 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. MethodsThe 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. ResultsThe 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). ConclusionThe DGHT has proven to be a general procedure, which can be easily applied to various tasks with high success rates. CARS, 2013 Object localization nationallicence Generalized Hough transform nationallicence Machine learning nationallicence Discriminative training nationallicence Image analysis nationallicence Automatic model training nationallicence Ruppertshofen Heike Department Digital Imaging, Philips Research Europe, Hamburg, Hamburg, Germany aut Lorenz Cristian Department Digital Imaging, Philips Research Europe, Hamburg, Hamburg, Germany aut Rose Georg Institute of Electronics, Signal Processing and Communication Technology, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany aut Schramm Hauke Institute of Applied Computer Science, University of Applied Sciences Kiel, Kiel, Germany aut International Journal of Computer Assisted Radiology and Surgery Springer Berlin Heidelberg 8/4(2013-07-01), 593-606 1861-6410 8:4<593 2013 8 11548 https://doi.org/10.1007/s11548-013-0817-7 text/html Onlinezugriff via DOI 1 research-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/s11548-013-0817-7 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Ruppertshofen Heike Department Digital Imaging, Philips Research Europe, Hamburg, Hamburg, Germany aut NATIONALLICENCE 700 1- Lorenz Cristian Department Digital Imaging, Philips Research Europe, Hamburg, Hamburg, Germany aut NATIONALLICENCE 700 1- Rose Georg Institute of Electronics, Signal Processing and Communication Technology, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany aut NATIONALLICENCE 700 1- Schramm Hauke Institute of Applied Computer Science, University of Applied Sciences Kiel, Kiel, Germany aut NATIONALLICENCE 773 0- International Journal of Computer Assisted Radiology and Surgery Springer Berlin Heidelberg 8/4(2013-07-01), 593-606 1861-6410 8:4<593 2013 8 11548 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer