# UICM-SOFF

**Repository Path**: LHY-CODE/UICM-SOFF

## Basic Information

- **Project Name**: UICM-SOFF
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2024-11-09
- **Last Updated**: 2026-01-30

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

This UICM-SOFF code for underwater image captioning task.

- **Python**: 3.6.9
- **Pytorch**: 1.7.1
- **Torchvision**: 0.8.2

We have publicly released the underwater image captioning dataset, which can be accessed and downloaded from the [UICD(underwater image captioning dataset)](https://gitee.com/LHY-CODE/UICD).

The degraded images of [ImageNet-100](https://aistudio.baidu.com/datasetdetail/208729) dataset can be generated using the provided code within this repository to train CNN: [Metalantis_UIE](https://gitee.com/wanghaoupc/Metalantis_UIE).

Object detection results for underwater images is extracted using the YOLOv8 implementation, the code can be obtained from: https://github.com/bubbliiiing/yolov8-pytorch.

After obtaining related preprocessed data can `train.py` and `test.py`.

**Acknowledgments**

- [https://github.com/ruotianluo/self-critical.pytorch](https://github.com/ruotianluo/self-critical.pytorch)
- [https://github.com/zchoi/S2-Transformer](https://github.com/zchoi/S2-Transformer)
- [https://github.com/NovaMind-Z/PTSN](https://github.com/NovaMind-Z/PTSN)