Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,64 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
---
|
4 |
+
|
5 |
+
# CoreML Image Colorizer Models
|
6 |
+
|
7 |
+
This repository contains CoreML versions of the ECCV16Colorizer and SIGGraph17Colorizer models, originally implemented in PyTorch. These models can be used to colorize black and white images on iOS devices.
|
8 |
+
|
9 |
+
## Models
|
10 |
+
|
11 |
+
- `ECCV16Colorizer.mlpackage`: The main CoreML model for image colorization from 2016.
|
12 |
+
- `SIGGraph17Colorizer.mlpackage`: The CoreML model for image colorization from 2017.
|
13 |
+
|
14 |
+
## Description
|
15 |
+
|
16 |
+
These CoreML models are converted from the PyTorch implementation of the ECCV16Colorizer. They use the 'L' channel of an image in the Lab color space to predict the 'a' and 'b' color channels, effectively colorizing black and white images.
|
17 |
+
|
18 |
+
## Usage
|
19 |
+
|
20 |
+
To use these models in your iOS app:
|
21 |
+
|
22 |
+
1. Download the desired `.mlpackage` file.
|
23 |
+
2. Add the model to your Xcode project.
|
24 |
+
3. Use CoreML framework to integrate the model into your app.
|
25 |
+
|
26 |
+
For a complete implementation example, check out the [GitHub repository](https://github.com/c2p-cmd/b-w_colorization).
|
27 |
+
|
28 |
+
## Model Input
|
29 |
+
|
30 |
+
- Format: MLShapedArray<Float32>
|
31 |
+
- Shape: [1, 1, 512, 512]
|
32 |
+
- Description: 'L' channel of the input image in Lab color space
|
33 |
+
|
34 |
+
## Model Output
|
35 |
+
|
36 |
+
- Format: MLShapedArray<Float32>
|
37 |
+
- Shape: [1, 2, 512, 512]
|
38 |
+
- Description: Predicted 'a' and 'b' channels in Lab color space
|
39 |
+
|
40 |
+
## Original Implementation
|
41 |
+
|
42 |
+
This CoreML model is based on the PyTorch implementation of ECCV16Colorizer. For more information about the original model and its implementation, please refer to the [original repository](https://github.com/richzhang/colorization).
|
43 |
+
|
44 |
+
## iOS Demo App
|
45 |
+
|
46 |
+
For a complete iOS app showcasing this model in action, check out our [GitHub repository](https://github.com/c2p-cmd/b-w_colorization).
|
47 |
+
|
48 |
+
## Citation
|
49 |
+
|
50 |
+
If you use this model in your research or project, please cite the original work:
|
51 |
+
|
52 |
+
```
|
53 |
+
@inproceedings{zhang2016colorful,
|
54 |
+
title={Colorful Image Colorization},
|
55 |
+
author={Zhang, Richard and Isola, Phillip and Efros, Alexei A},
|
56 |
+
booktitle={ECCV},
|
57 |
+
year={2016}
|
58 |
+
}
|
59 |
+
```
|
60 |
+
|
61 |
+
---
|
62 |
+
|
63 |
+
Created by Sharan Thakur
|
64 |
+
GitHub: [https://github.com/c2p-cmd](https://github.com/c2p-cmd)
|