Update README.md
Browse files
README.md
CHANGED
|
@@ -10,28 +10,31 @@ datasets:
|
|
| 10 |
- ModelMoe/DDRD20K
|
| 11 |
---
|
| 12 |
|
| 13 |
-
# DDRM: Diffusion Degradation Restoration Model
|
| 14 |
|
| 15 |
-
|
|
|
|
| 16 |
|
| 17 |
-
|
| 18 |
|
| 19 |
-
|
| 20 |
-
* 跨图像语义对齐:模型能够从参考图中提取高质量的色彩先验。
|
| 21 |
-
* 查询向量机制 (Queries):利用 100 个可学习的查询向量在参考图特征空间中进行“搜索”,自动拾取最契合当前场景的色彩与光影参数。
|
| 22 |
|
| 23 |
-
|
| 24 |
-
* 模型内置了通过核密度估计 (Kernel Density Estimation) 实时捕捉图像的统计特性。
|
| 25 |
-
* 全局色彩锚点:模型不再仅仅依赖像素值,而是通过分析图像的 PDF (概率密度函数) 特征,确保还原后的图像在全局色调上与参考图保持统计学一致。
|
| 26 |
|
| 27 |
-
|
| 28 |
-
*
|
| 29 |
-
*
|
| 30 |
|
| 31 |
-
##
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
|
|
|
|
|
|
|
|
|
| 35 |
|
| 36 |
-
|
| 37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
- ModelMoe/DDRD20K
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# DDRM: Diffusion Degradation Restoration Model / 扩散退化还原模型
|
| 14 |
|
| 15 |
+
## Model Description / 模型简介
|
| 16 |
+
**DDRM** (Diffusion Degradation Restoration Model) is a deep learning framework engineered for high-precision image restoration and color alignment. By leveraging **Reference-based Color Querying**, the model achieves accurate restoration of complex degraded images based on external stylistic cues.
|
| 17 |
|
| 18 |
+
**DDRM** 是一款专为高精度图像修复与色彩对齐设计的深度学习模型。该模型通过**参考图颜色查询**机制,实现了对复杂退化图像的精准还原与色调同步。
|
| 19 |
|
| 20 |
+
---
|
|
|
|
|
|
|
| 21 |
|
| 22 |
+
## 🛠️ Model Features / 模型架构特性
|
|
|
|
|
|
|
| 23 |
|
| 24 |
+
### 1. Reference-based Color Query / 交互式参考图颜色查询
|
| 25 |
+
* **Cross-Image Semantic Alignment:** Extracts high-quality color priors from reference images. / **跨图像语义对齐**:从参考图中提取高质量的色彩先验信息。
|
| 26 |
+
* **Learnable Query Mechanism:** Utilizes 100 learnable query vectors to "search" the reference feature space, automatically selecting the optimal color and lighting parameters for the target scene. / **查询向量机制**:利用 100 个可学习的查询向量在参考图特征空间中进行“搜索”,自动拾取最契合当前场景的色彩与光影参数。
|
| 27 |
|
| 28 |
+
### 2. KDE-based Distribution Prior / 动态核密度估计先验
|
| 29 |
+
* **Statistical Characteristic Capture:** Built-in Kernel Density Estimation (KDE) captures real-time image statistical properties. / **统计特性捕捉**:内置核密度估计模块,实时捕捉图像的统计学特征。
|
| 30 |
+
* **Global Color Anchors:** Instead of relying solely on pixel values, the model analyzes the PDF (Probability Density Function) to ensure the restored image maintains global tonal consistency with the reference. / **全局色彩锚点**:模型不再仅仅依赖像素值,而是通过分析概率密度函数 (PDF) 特征,确保还原图像在全局色调上与参考图保持统计学一致。
|
| 31 |
+
|
| 32 |
+
### 3. Multi-Scale Transformer Decoder / 多尺度 Transformer 解码架构
|
| 33 |
+
* **Hybrid Architecture:** Combines the local fidelity of **ConvNeXt V2** with the global receptive field of **Transformers**. / **混合架构**:结合了 ConvNeXt V2 的局部保真度与 Transformer 的全局感受野。
|
| 34 |
+
- **Decoupled Representation Learning:** The encoder extracts structure and texture, while the multi-scale Transformer decoder injects color information guided by the Query mechanism. / **解耦表征学习**:编码器负责提取结构与纹理,而多尺度 Transformer 解码器负责注入由 Query 引导的色彩信息。
|
| 35 |
|
| 36 |
+
---
|
| 37 |
+
|
| 38 |
+
## 🚀 Getting Started / 环境依赖
|
| 39 |
+
```bash
|
| 40 |
+
pip install torch kornia timm opencv-python
|