Computer vision is continuously advancing in the field of accurate visual representation. Deep learning architectures such as ResNet50 and pooling algorithms have been crucial to this process. In this ...

To solve the above issues, we propose a local structure visualization descriptor to preserve the original structure information, and cast the feature matching task into an evaluation of the consensus ...