The effects of different levels of realism on the training of CNNs with only synthetic images for the semantic segmentation of robotic instruments in a head phantom.
Saúl Alexis Heredia PérezMurilo Marques MarinhoKanako HaradaMamoru MitsuishiPublished in: Int. J. Comput. Assist. Radiol. Surg. (2020)
Keyphrases
- semantic segmentation
- street scenes
- label transfer
- conditional random fields
- superpixels
- weakly supervised
- scene classification
- three dimensional
- pascal voc
- object recognition
- information extraction
- bag of words
- object categories
- image set
- long range
- learning algorithm
- probabilistic model
- training set
- feature selection