70 lines
1.3 KiB
Markdown
70 lines
1.3 KiB
Markdown
|
# 激活虚拟环境
|
|||
|
|
.\.venv\Scripts\activate
|
|||
|
|
|
|||
|
|
# 确保数据集已正确放置在data/目录下
|
|||
|
|
# 验证数据集结构：
|
|||
|
|
# - data/images/ (原始图像)
|
|||
|
|
# - data/masks/ (对应掩码)
|
|||
|
|
|
|||
|
|
# 运行UNet训练
|
|||
|
|
python code/unet_segmentation.py
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
预期输出：
|
|||
|
|
```
|
|||
|
|
Epoch [1/10], Loss: 0.4567
|
|||
|
|
Epoch [2/10], Loss: 0.3214
|
|||
|
|
...
|
|||
|
|
训练完成，总耗时: 125.36 秒
|
|||
|
|
模型已保存为 unet_brain_tumor.pth
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
## 运行图像分割实验（COCO格式数据集）
|
|||
|
|
```bash
|
|||
|
|
# 激活虚拟环境
|
|||
|
|
.\.venv\Scripts\activate
|
|||
|
|
|
|||
|
|
# 确保数据集已正确放置在dataset/目录下
|
|||
|
|
# 验证数据集结构：
|
|||
|
|
# - dataset/train/ (训练集图像)
|
|||
|
|
# - dataset/valid/ (验证集图像)
|
|||
|
|
# - dataset/test/ (测试集图像)
|
|||
|
|
# - _annotations.coco.json (标注文件)
|
|||
|
|
|
|||
|
|
# 运行UNet训练
|
|||
|
|
python code/unet_segmentation.py
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
预期输出：
|
|||
|
|
```
|
|||
|
|
INFO - 成功加载 N 个训练样本
|
|||
|
|
INFO - Epoch [1/10], Loss: 0.4567
|
|||
|
|
INFO - Epoch [2/10], Loss: 0.3214
|
|||
|
|
...
|
|||
|
|
INFO - 训练完成，总耗时: 125.36 秒
|
|||
|
|
INFO - 模型已保存为 unet_coco.pth```
|
|||
|
|
|
|||
|
|
## 运行SIFT特征提取实验
|
|||
|
|
```bash
|
|||
|
|
# 激活虚拟环境
|
|||
|
|
.\.venv\Scripts\activate
|
|||
|
|
|
|||
|
|
# 运行SIFT特征提取
|
|||
|
|
python code/sift_features.py
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
预期输出：
|
|||
|
|
```
|
|||
|
|
开始SIFT特征提取...
|
|||
|
|
检测到 128 个关键点
|
|||
|
|
描述符形状: (128, 128)
|
|||
|
|
特征提取耗时: 0.0456 秒
|
|||
|
|
结果已保存至: output/sift_results/sift_result.jpg```
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|
|||
|
|
|
|||
|
|
```
|