Semantic Segmentation of Fish and Underwater Environments Using Deep Convolutional Neural Networks and Learned Active Contours.
Miguel ChicchónHector BedonCarlos R. del-BlancoIvan SipiranPublished in: IEEE Access (2023)
Keyphrases
- active contours
- semantic segmentation
- convolutional neural networks
- superpixels
- segmentation method
- object boundaries
- conditional random fields
- level set
- scene classification
- weakly supervised
- shape prior
- image segmentation
- energy function
- object class
- object categories
- object recognition
- object segmentation
- multiscale
- object classes
- coarse to fine
- graph structure
- natural images
- semi supervised
- machine learning