S3G is the implementation module for the oil spill superpixel segmentation function mentioned in relevant papers. It is a superpixel generation algorithm based on the SSQM (the Social Support Quantitative Model), using SAM (Spectral Angle Mapper) as the pixel similarity measurement standard and supporting multi parameter input.
The code consists of three parts：
- Main.m: The main file of S3G, including data reading, superpixel boundaries, result generation, and storage functions.
- S3G.m: S3G function file, designed to implement superpixel segmentation based on S3G.
- UEE_evaluation: Evaluation file, used to evaluate the results of S3G based superpixel segmentation. The indicator used is UEE.
The second stage semantic segmentation in the paper is based on the DeepLabV3+ algorithm (Open source acquisition), which adds superpixel segmentation results as an auxiliary and does not involve original innovation or improvement. Therefore, it is inconvenient to upload it to this repository.
This algorithm is a functional module used in relevant papers, and detailed content and functions can be found in the paper.