FeDNet: Feature Decoupled Network for polyp segmentation from endoscopy images.
Yanzhou SuJian ChengChuqiao ZhongYijie ZhangJin YeJunjun HeJun LiuPublished in: Biomed. Signal Process. Control. (2023)
Keyphrases
- segmentation method
- segmentation algorithm
- image features
- test images
- image analysis
- microscopic images
- segmented images
- image regions
- ground truth
- pixel wise
- adaptive thresholding
- grey level
- edge detection
- segmented regions
- tubular structures
- piece wise
- image database
- image data
- image retrieval
- object recognition
- accurate segmentation
- gray value
- region growing
- fully automatic
- automatically segmented
- geometrical features
- segmentation errors
- image classification
- input image
- unsupervised segmentation
- segmentation accuracy
- microscopy images
- low depth of field
- three dimensional
- multiple objects
- shape prior
- image segmentation
- color and texture information
- image slices
- image registration
- positron emission tomography
- level set
- image segments
- video images
- object segmentation
- bounding box
- cell nuclei
- medical images
- low contrast
- feature vectors
- textural features
- intensity images
- feature maps