Leveraging Random Forests for Interactive Exploration of Large Histological Images

Abstract : Thelargesizeofhistologicalimagescombinedwiththeirvery challenging appearance are two main difficulties which considerably com- plicate their analysis. In this paper, we introduce an interactive strategy leveraging the output of a supervised random forest classifier to guide a user through such large visual data. Starting from a forest-based pixel- wise estimate, subregions of the images at hand are automatically ranked and sequentially displayed according to their expected interest. After each region suggestion, the user selects among several options a rough es- timate of the true amount of foreground pixels in this region. From these one-click inputs, the region scoring function is updated in real time using an online gradient descent procedure, which corrects on-the-fly the short- comings of the initial model and adapts future suggestions accordingly. Experimental validation is conducted for extramedullary hematopoesis localization and demonstrates the practical feasibility of the procedure as well as the benefit of the online adaptation strategy.
Type de document :
Communication dans un congrès
Int. Conf. on Medical Image Computing and Computer Assisted Intervention, MICCAI 2014, Sep 2014, Boston, United States. 2014
Liste complète des métadonnées

Littérature citée [12 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-01056993
Contributeur : Eric Marchand <>
Soumis le : jeudi 21 août 2014 - 10:16:13
Dernière modification le : mardi 16 janvier 2018 - 15:54:11
Document(s) archivé(s) le : jeudi 27 novembre 2014 - 12:35:24

Fichier

peter2014miccai.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01056993, version 1

Citation

Loic Peter, Diana Mateus, Pierre Chatelain, Noemi Schworn, Stefan Stangl, et al.. Leveraging Random Forests for Interactive Exploration of Large Histological Images. Int. Conf. on Medical Image Computing and Computer Assisted Intervention, MICCAI 2014, Sep 2014, Boston, United States. 2014. 〈hal-01056993〉

Partager

Métriques

Consultations de la notice

278

Téléchargements de fichiers

155