Jiafeng Huang1, Tianjun Zhang1, Shengjie Zhao1, Lin Zhang1, and Yicong Zhou2
1School of Software Engineering, Tongji University, Shanghai, China
2Department of Computer and Information Science, University of Macau, China
This is the website for our paper An Annotated Underwater Organism Image Benchmark Dataset and A Lightweight Module Designed for the Object Detection Networks.
MKUO is the underwater biological data set covering the broadest categories and the first one labeled with taxonomy scientific name.
In each column, the image without annotations is shown on the top, and the corresponding result with annotations is shown at the bottom. From left to right, the samples of Pneumatophorus japonicus, Aurelia aurita, Scarus ghobban and Echeneis naucrates are given, respectively.
Considering all metrics comprehensively, the lightweight effect of the Sparse Ghost Module is the best among all compared schemes.