|
1 | | -# NeuralNetworks |
2 | | -[](https://github.com/psf/black) |
| 1 | +<p align="center"> |
| 2 | + <a href="https://github.com/ContentAutomation"><img src="https://contentautomation.s3.eu-central-1.amazonaws.com/logo.png" alt="Logo" width="150"/></a> |
| 3 | + <br /> |
| 4 | + <br /> |
| 5 | + <a href="http://choosealicense.com/licenses/mit/"><img src="https://img.shields.io/badge/license-MIT-3C93B4.svg?style=flat" alt="MIT License"></a> |
| 6 | + <a href="https://github.com/psf/black"><img src="https://img.shields.io/badge/code%20style-black-000000.svg" alt="Code style: black"></a> |
| 7 | + <br /> |
| 8 | + <a href="https://github.com/tensorflow/tensorflow"><img src="https://img.shields.io/badge/TensorFlow-FF6F00.svg?logo=tensorflow&logoColor=white" alt="Tensorflow"></a> |
| 9 | + <br /> |
| 10 | + <br /> |
| 11 | + <i>Implementation of a Neural Network that can detect whether a video is Ingame or not </i> |
| 12 | + <br /> |
| 13 | +<br /> |
| 14 | + <i><b>Authors</b>: |
| 15 | + <a href="https://github.com/ChristianCoenen">Christian C.</a>, |
| 16 | + <a href="https://github.com/MorMund">Moritz M.</a>, |
| 17 | + <a href="https://github.com/lucaSchilling">Luca S.</a> |
| 18 | + </i> |
| 19 | +</p> |
| 20 | +<hr /> |
| 21 | + |
| 22 | +<p align="center"> |
| 23 | + <img src="https://contentautomation.s3.eu-central-1.amazonaws.com/detect_ingame.png" alt="Ingame Detection" width="1200"/></a> |
| 24 | +</p> |
| 25 | + |
| 26 | + |
| 27 | +## Setup |
| 28 | +This project requires [Poetry](https://python-poetry.org/) to install the required dependencies. |
| 29 | +Check out [this link](https://python-poetry.org/docs/) to install Poetry on your operating system. |
| 30 | + |
| 31 | +Make sure you have installed [Python](https://www.python.org/downloads/) 3.8 or higher! Otherwise Step 3 will let you know that you have no compatible Python version installed. |
| 32 | + |
| 33 | +1. Clone/Download this repository |
| 34 | + |
| 35 | + *Note: For test models/assets, download [Release v1.0](https://github.com/ContentAutomation/NeuralNetworks/releases/latest)* |
| 36 | +2. Navigate to the root of the repository |
| 37 | +3. Run ```poetry install``` to create a virtual environment with Poetry |
| 38 | +4. Run ```poetry run python src/filename.py``` to run the program. Alternatively you can run ```poetry shell``` followed by ```python src/filename.py``` |
| 39 | +5. Enjoy :) |
| 40 | + |
| 41 | +## Script Explanations |
| 42 | + |
| 43 | +### video2images.py |
| 44 | +This utility can be used to build the dataset by splitting video files into images. |
| 45 | + |
| 46 | +### predict.py |
| 47 | +This script is used to verify the performance of the trained neural network |
| 48 | +by specifying a path to the model of the trained neural network, |
| 49 | +and a video clip that should be analyzed. |
| 50 | + |
| 51 | +### game_detection.py |
| 52 | +This script is used to train a neural network (e.g. create a model) on a given dataset. |
| 53 | +If enough data is present, the neural network will learn to distinguish Ingame clips from clips |
| 54 | +that are not ingame (e.g. Lobby, Queue, ...) |
| 55 | + |
| 56 | +## Creating a new model for a game |
| 57 | +Let's assume you want to create a new model for the game Dota2. The following steps have to be performed: |
| 58 | +1. Download clips for Dota2 that are both ingame and not ingame (recommended source: Twitch) |
| 59 | + |
| 60 | + HINT: You can download clips manually or by creating a compilation with [TwitchCompilationCreator](https://github.com/ContentAutomation/TwitchCompilationCreator) |
| 61 | + |
| 62 | +2. Split the clips into images via ```video2images.py``` |
| 63 | +3. Create the following folder structure |
| 64 | +```bash |
| 65 | +... |
| 66 | +│ |
| 67 | +└───anyFolderName |
| 68 | + │ |
| 69 | + └───dota2 |
| 70 | + └───nogame |
| 71 | +``` |
| 72 | +4. Sort the clips from step 1 into those folders depending on if they are ingame or not |
| 73 | +5. Create a ```main.py``` file in ```./src/``` to initialize a ```GameDetection``` object, then run it (see example below) |
| 74 | +6. Test the created model on a few example clips using ```predict.py``` to verify its accuracy |
| 75 | + |
| 76 | + NOTE: The number of images in the 'gamename' or 'nogame' folder has to be greater than or equal to the defined batch size |
| 77 | +```python |
| 78 | +# For more information about the parameters, check out game_detection.py |
| 79 | +m = GameDetection( |
| 80 | + model_name="ResNet50", |
| 81 | + game_name="dota2", |
| 82 | + dataset_path="---PATH TO 'anyFolderName'---", |
| 83 | + input_size=(224, 224), |
| 84 | + batch_size=16, |
| 85 | + save_generated_images=False, |
| 86 | + convert_to_gray=False, |
| 87 | +) |
| 88 | +m.train(epochs=2) |
| 89 | +``` |
0 commit comments