Towards Smaller and Stronger: An Edge-Aware Lightweight Segmentation Approach for Unmanned Surface Vehicles in Water Scenarios

Abstract

The accurate detection and segmentation of accessible surface regions in water scenarios is one of the indispensable capabilities of surface unmanned vehicle systems. ‘Most existing methods focus on accuracy and ignore the lightweight and real-time demands. Therefore, they are not suitable for embedded devices, which have been wildly applied in practical applications.‘ An edge-aware lightweight water scenario segmentation method (ELNet), which establishes a lighter yet better network with lower computation, is proposed. ELNet utilizes two-stream learning and edge-prior information. Except for the context stream, a spatial stream is expanded to learn spatial details in low-level layers with no extra computation cost in the inference stage. Meanwhile, edge-prior information is introduced to the two streams, which expands the perspectives of pixel-level visual modeling. The experimental results are 45.21% in FPS, 98.5% in detection robustness, 75.1% in F-score on MODS benchmark, 97.82% in precision, and 93.96% in F-score on USV Inland dataset. It demonstrates that ELNet uses fewer parameters to achieve comparable accuracy and better real-time performance.

Fast-MFQE: A Fast Approach for Multi-Frame Quality Enhancement on Compressed Video

“…Nevertheless, it is challenging to apply these methods to other compression tasks due to the limited generalization ability. With the advance of the deep learning method, an expanding range of methods have embraced convolutional neural network (CNN) approaches [ 12 , 13 , 14 , 15 , 16 ] to improve the compressed image quality. In [ 12 ], a four-layer AR-CNN was first introduced to deal with various artifacts in JPEG images.…”

“…The term “lightweight model” refers to compressing the model size to maximize computational speed while preserving the accuracy. Researchers have been paying increasing attention to developing lightweight models in the field of image classification to enable deployment on mobile devices [ 14 , 15 , 16 , 27 , 28 ]. Among the pioneer endeavors in developing lightweight models, SqueezeNet [ 29 ] emerged, replacing 3 × 3 convolutions with 1 × 1 convolutions, resulting in a parameter reduction of approximately one-fiftieth compared to AlexNet [ 30 ].…”