SECODA: Segmentation- and Combination-Based Detection of Anomalies.
Ralph FoorthuisPublished in: CoRR (2020)
Keyphrases
- anomaly detection
- object detection
- fully automatic
- image segmentation
- automatic detection
- reliable detection
- segmentation method
- detection method
- landmark detection
- detection algorithm
- cell segmentation
- level set
- complex background
- multiscale
- false alarms
- low signal to noise ratio
- lesion segmentation
- video segmentation
- region growing
- false positives
- rib cage
- boundary detection
- segmented images
- segmentation accuracy
- automated analysis
- bounding box
- event detection
- segmentation algorithm
- image analysis
- abnormal behavior
- computed tomography angiography
- circular hough transform
- threshold selection
- detection rate
- object segmentation