Abstract
Autonomous ships rely on sensory data to perceive objects of interest in their environment. Deep Learning based object detection in the image domain commonly used to solve this issue. The robustness of such approaches in non-ideal conditions is, however, still to be proven. In this work state of the art methods are applied on the RetinaNet architecture attempting to create a more robust object detection network given noisy input data. The GroupSort activation function and Spectral Normalization is used and the results are compared to the standard RetinaNet network. Our findings show that these modifications perform better and ensure robustness under moderate noise levels, than the standard RetinaNet network.
Highlights
Traversing the ocean can present many challenges to an autonomous vessel
Our findings show that these modifications perform better and ensure robustness under moderate noise levels, than the standard RetinaNet network
The goal of this paper is to explore architectures that add robustness to object detection against noisy input in marine environments
Summary
Traversing the ocean can present many challenges to an autonomous vessel. A diverse set of weather conditions, equipment malfunctioning, continuous movement all need to be taken into account. Given a neural network, very slight changes to the input can have compounding effects on each layer and result in very different output. GroupSort, as introduced in [1], is a Lipschitz continuous activation function, it sorts the pre-activated weights into some selected grouping. For example a GroupSort with a grouping size of 2, will split the weights into groups of 2, sort the weights in the groups, as can be seen done in figure 3. This operation is a nonlinear operation which is differentiable as the Jacobian is a permutation matrix which preserve every p-norm [1].
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IOP Conference Series: Materials Science and Engineering
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
