Can Masses of Non-Experts Train Highly Accurate Image Classifiers? - A Crowdsourcing Approach to Instrument Segmentation in Laparoscopic Images.
Lena Maier-HeinSven MersmannDaniel KondermannSebastian BodenstedtAlexandro SanchezChristian StockHannes Götz KenngottMatthias EisenmannStefanie SpeidelPublished in: MICCAI (2) (2014)
Keyphrases
- highly accurate
- test images
- segmentation method
- input image
- segmentation algorithm
- image regions
- grey level
- segmented images
- image pixels
- image data
- image analysis
- image collections
- edge detection
- image segments
- adaptive thresholding
- microscopic images
- image features
- image slices
- low depth of field
- gray level images
- gray value
- brain mr images
- image segmentation
- images of natural scenes
- pixel wise
- pixel level
- image segmentation algorithm
- image retrieval
- segmentation errors
- homogeneous regions
- segmentation accuracy
- microscopy images
- image classification
- image content
- berkeley segmentation dataset
- automatically segmented
- intensity distribution
- image segmentation algorithms
- pixel values
- intensity images
- segmented regions
- region of interest
- foreground and background
- fundamental problem in computer vision
- breast ultrasound
- bounding box
- multiple objects
- segmentation scheme
- capable of producing
- microscope images
- multiscale
- circular hough transform
- intensity variations
- image set
- object segmentation
- feature points
- shape prior
- region growing
- labeled images
- high quality
- gray level
- color features
- region segmentation
- colour images
- contrast enhancement
- medical imaging
- single image
- active contours
- watershed segmentation
- high accuracy
- image registration
- texture and shape features
- image structure
- level set