Zero-shot incremental learning aims to enable a model to generalize to new classes without forgetting previously learned classes. However, the semantic gap between old and new sample classes can lead ...
The following are the summary of the paper’s contributions: A self-adaptive dual approach physics guided image enhancement module for restoring underwater images. An improved attention-based ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results