%0 Conference Paper %F Oral %T Deeply optimized Hough transform: Application to action segmentation %+ Département Intelligence Ambiante et Systèmes Interactifs (DIASI) %+ Institut des Systèmes Intelligents et de Robotique (ISIR) %A Chan-Hon-Tong, A. %A Achard, C. %A Lucat, L. %Z Conference of 17th International Conference on Image Analysis and Processing, ICIAP 2013 ; Conference Date: 9 September 2013 Through 13 September 2013; Conference Code:99647 %< avec comité de lecture %B 17th International Conference on Image Analysis and Processing, ICIAP 2013 %C Naples, Italy %S Image Analysis and Processing – ICIAP 2013. Lecture Notes in Computer Science %V 8156 LNCS %N PART 1 %P 51-60 %8 2013-09-09 %D 2013 %R 10.1007/978-3-642-41181-6_6 %K Optimization %K Motion estimation %K Image analysis %K Skeleton extraction %K Object Detection %K Multiple computers %K Hough transforms %K Human action detections %K Implicit shape models %K Linear programming solvers %K Hough Transform %K Learning %K Action Segmentation %Z Computer Science [cs]Conference papers %X Hough-like methods like Implicit Shape Model (ISM) and Hough forest have been successfully applied in multiple computer vision fields like object detection, tracking, skeleton extraction or human action detection. However, these methods are known to generate false positives. To handle this issue, several works like Max-Margin Hough Transform (MMHT) or Implicit Shape Kernel (ISK) have reported significant performance improvements by adding discriminative parameters to the generative ones introduced by ISM. In this paper, we offer to use only discriminative parameters that are globally optimized according to all the variables of the Hough transform. To this end, we abstract the common vote process of all Hough methods into linear equations, leading to a training formulation that can be solved using linear programming solvers. Our new Hough Transform significantly outperforms the previous ones on HoneyBee and TUM datasets, two public databases of action and behaviour segmentation. %G English %2 https://cea.hal.science/cea-01813733/document %2 https://cea.hal.science/cea-01813733/file/Chan-Hon-Tong2013.pdf %L cea-01813733 %U https://cea.hal.science/cea-01813733 %~ CEA %~ UPMC %~ CNRS %~ ISIR %~ DRT %~ UPMC_POLE_1 %~ LIST %~ SORBONNE-UNIVERSITE %~ SU-SCIENCES %~ SU-TI %~ ISIR_PIROS %~ ALLIANCE-SU