Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,150 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- gamedev
|
| 5 |
+
- normal-map
|
| 6 |
+
- image-to-image
|
| 7 |
+
- computer-vision
|
| 8 |
+
- sprites
|
| 9 |
+
- pixel-art
|
| 10 |
+
- onnx
|
| 11 |
+
- 2d-lighting
|
| 12 |
+
pipeline_tag: image-to-image
|
| 13 |
+
library_name: onnx
|
| 14 |
---
|
| 15 |
+
|
| 16 |
+
<div align="center">
|
| 17 |
+
|
| 18 |
+
# 🎮 Sprite2Normal
|
| 19 |
+
|
| 20 |
+
**AI-powered Tangent-Space Normal Map Generator for 2D Sprites & Pixel Art**
|
| 21 |
+
|
| 22 |
+
[ 🇬🇧 English ](#-english) • [ 🇷🇺 Русский ](#-русский)
|
| 23 |
+
|
| 24 |
+
</div>
|
| 25 |
+
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+
<a name="-english"></a>
|
| 29 |
+
## 🇬🇧 English
|
| 30 |
+
|
| 31 |
+
**Sprite2Normal** is an ultra-lightweight neural network (~7 MB) designed to instantly generate tangent-space **Normal Maps** from flat 2D sprites, textures, and pixel art.
|
| 32 |
+
|
| 33 |
+
Built for indie game developers to easily add dynamic 2D lighting to their games without manually drawing height or normal maps.
|
| 34 |
+
|
| 35 |
+
Trained completely from scratch with a custom lightweight U-Net architecture.
|
| 36 |
+
|
| 37 |
+
### ✨ Features
|
| 38 |
+
- **Ultra-Compact:** Model file is only **~7 MB** in ONNX format.
|
| 39 |
+
- **CPU & GPU Friendly:** Runs in **3–10 ms** on standard CPUs without requiring a dedicated GPU.
|
| 40 |
+
- **Engine Compatible:** Generates standard OpenGL tangent-space normal maps (X: Red, Y: Green, Z: Blue), ready for **Godot, Unity, GameMaker, Unreal Engine**, and custom shaders.
|
| 41 |
+
- **Dynamic Resolution:** Fully convolutional architecture supporting sprites of any size.
|
| 42 |
+
|
| 43 |
+
### 🚀 Quick Start (Python)
|
| 44 |
+
|
| 45 |
+
#### 1. Install dependencies
|
| 46 |
+
```bash
|
| 47 |
+
pip install onnxruntime pillow numpy
|
| 48 |
+
```
|
| 49 |
+
|
| 50 |
+
#### 2. Inference script
|
| 51 |
+
```python
|
| 52 |
+
import numpy as np
|
| 53 |
+
from PIL import Image
|
| 54 |
+
import onnxruntime as ort
|
| 55 |
+
|
| 56 |
+
session = ort.InferenceSession("sprite2normal.onnx", providers=["CPUExecutionProvider"])
|
| 57 |
+
input_name = session.get_inputs()[0].name
|
| 58 |
+
output_name = session.get_outputs()[0].name
|
| 59 |
+
|
| 60 |
+
img = Image.open("character.png").convert("RGB")
|
| 61 |
+
orig_w, orig_h = img.size
|
| 62 |
+
|
| 63 |
+
img_resized = img.resize((128, 128), Image.Resampling.BILINEAR)
|
| 64 |
+
arr = np.array(img_resized, dtype=np.float32) / 255.0
|
| 65 |
+
arr = np.transpose(arr, (2, 0, 1))
|
| 66 |
+
input_tensor = np.expand_dims(arr, axis=0)
|
| 67 |
+
|
| 68 |
+
output = session.run([output_name], {input_name: input_tensor})[0][0]
|
| 69 |
+
|
| 70 |
+
pred = (output + 1.0) * 0.5 * 255.0
|
| 71 |
+
pred = np.clip(pred, 0, 255).astype(np.uint8)
|
| 72 |
+
pred = np.transpose(pred, (1, 2, 0))
|
| 73 |
+
|
| 74 |
+
normal_map = Image.fromarray(pred).resize((orig_w, orig_h), Image.Resampling.BILINEAR)
|
| 75 |
+
normal_map.save("character_normal.png")
|
| 76 |
+
print("Saved to character_normal.png!")
|
| 77 |
+
```
|
| 78 |
+
|
| 79 |
+
### 🕹️ Game Engine Integration
|
| 80 |
+
- **Godot Engine:** Place `character.png` and `character_normal.png` in your project folder. Assign the normal map to the **Normal Map** slot of your `Sprite2D` node and add a `PointLight2D`.
|
| 81 |
+
- **Unity (URP 2D):** Set texture type of `character_normal.png` to **Normal Map**. Assign it to **Secondary Textures** of your Sprite Renderer.
|
| 82 |
+
|
| 83 |
+
---
|
| 84 |
+
|
| 85 |
+
<a name="-русский"></a>
|
| 86 |
+
## 🇷🇺 Русский
|
| 87 |
+
|
| 88 |
+
<details open>
|
| 89 |
+
<summary><b>Нажмите, чтобы свернуть / развернуть русский текст</b></summary>
|
| 90 |
+
<br>
|
| 91 |
+
|
| 92 |
+
**Sprite2Normal** — это сверхлёгкая нейросеть (~7 МБ), предназначенная для мгновенной генерации карт нормалей (**Normal Maps**) из плоских 2D-спрайтов, текстур и пиксель-арта.
|
| 93 |
+
|
| 94 |
+
Создана специально для инди-разработчиков игр, чтобы быстро добавлять динамическое 2D-освещение без необходимости вручную прорисовывать объём каждого кадра анимации.
|
| 95 |
+
|
| 96 |
+
Обучена полностью с нуля на специализированной компактной архитектуре U-Net.
|
| 97 |
+
|
| 98 |
+
### ✨ Особенности
|
| 99 |
+
- **Минимальный вес:** Модель весит всего **~7 МБ** в формате ONNX.
|
| 100 |
+
- **Работает без мощной видеокарты:** Время инференса всего **3–10 мс** на обычном офисном процессоре (CPU).
|
| 101 |
+
- **Готовность к игровым движкам:** Создаёт стандартные OpenGL-карты нормалей (X: Красный, Y: Зелёный, Z: Синий), совместимые с **Godot, Unity, GameMaker, Unreal Engine** и кастомными шейдерами.
|
| 102 |
+
- **Любое разрешение:** Свёрточная архитектура принимает спрайты произвольного размера.
|
| 103 |
+
|
| 104 |
+
### 🚀 Быстрый старт (Python)
|
| 105 |
+
|
| 106 |
+
#### 1. Установка библиотек
|
| 107 |
+
```bash
|
| 108 |
+
pip install onnxruntime pillow numpy
|
| 109 |
+
```
|
| 110 |
+
|
| 111 |
+
#### 2. Скрипт запуска
|
| 112 |
+
```python
|
| 113 |
+
import numpy as np
|
| 114 |
+
from PIL import Image
|
| 115 |
+
import onnxruntime as ort
|
| 116 |
+
|
| 117 |
+
session = ort.InferenceSession("sprite2normal.onnx", providers=["CPUExecutionProvider"])
|
| 118 |
+
input_name = session.get_inputs()[0].name
|
| 119 |
+
output_name = session.get_outputs()[0].name
|
| 120 |
+
|
| 121 |
+
img = Image.open("character.png").convert("RGB")
|
| 122 |
+
orig_w, orig_h = img.size
|
| 123 |
+
|
| 124 |
+
img_resized = img.resize((128, 128), Image.Resampling.BILINEAR)
|
| 125 |
+
arr = np.array(img_resized, dtype=np.float32) / 255.0
|
| 126 |
+
arr = np.transpose(arr, (2, 0, 1))
|
| 127 |
+
input_tensor = np.expand_dims(arr, axis=0)
|
| 128 |
+
|
| 129 |
+
output = session.run([output_name], {input_name: input_tensor})[0][0]
|
| 130 |
+
|
| 131 |
+
pred = (output + 1.0) * 0.5 * 255.0
|
| 132 |
+
pred = np.clip(pred, 0, 255).astype(np.uint8)
|
| 133 |
+
pred = np.transpose(pred, (1, 2, 0))
|
| 134 |
+
|
| 135 |
+
normal_map = Image.fromarray(pred).resize((orig_w, orig_h), Image.Resampling.BILINEAR)
|
| 136 |
+
normal_map.save("character_normal.png")
|
| 137 |
+
print("Сохранено в character_normal.png!")
|
| 138 |
+
```
|
| 139 |
+
|
| 140 |
+
### 🕹️ Интеграция в движки
|
| 141 |
+
- **Godot Engine:** Поместите спрайт и карту нормалей в проект. В ноде `Sprite2D` в поле **Normal Map** выберите созданный файл и добавьте `PointLight2D`.
|
| 142 |
+
- **Unity (URP 2D):** В настройках текстуры укажите тип **Normal Map** и назначьте её во вкладку **Secondary Textures** компонента Sprite Renderer.
|
| 143 |
+
|
| 144 |
+
</details>
|
| 145 |
+
|
| 146 |
+
---
|
| 147 |
+
|
| 148 |
+
## 📄 License / Лицензия
|
| 149 |
+
This project is open-sourced under the **Apache-2.0** License.
|
| 150 |
+
Проект распространяется под открытой лицензией **Apache-2.0**.
|