https://github.com/NikolasEnt/Lyft-Perception-Challenge

The 4th place and the fastest solution of the Lyft Perception Challenge (Image semantic segmentation with PyTorch)
https://github.com/NikolasEnt/Lyft-Perception-Challenge

computer-vision deep-learning linknet pytorch sdcnd self-driving-car semantic-segmentation

Last synced: 3 months ago
JSON representation

The 4th place and the fastest solution of the Lyft Perception Challenge (Image semantic segmentation with PyTorch)

• Host: GitHub
• URL: https://github.com/NikolasEnt/Lyft-Perception-Challenge
• Owner: NikolasEnt
• Created: 2018-05-09T12:57:41.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-08-23T22:59:49.000Z (about 6 years ago)
• Last Synced: 2023-11-02T03:31:42.367Z (about 1 year ago)
• Topics: computer-vision, deep-learning, linknet, pytorch, sdcnd, self-driving-car, semantic-segmentation
• Language: Jupyter Notebook
• Homepage: https://nikolasent.github.io/proj/comp2
• Size: 8.83 MB
• Stars: 68
• Watchers: 8
• Forks: 12
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
## The Lyft Perception Challenge

### Challenge Overview

![readme_img](readme_img/title.gif)

Your goal in this challenge is pixel-wise identification of objects in camera images. The segmentation targets are other cars and the drivable area of the road.

Images from the [CARLA](http://carla.org/) 0.8.2 simulator are used for training and testing in the challenge.

For details see my website: __[https://NikolasEnt.github.io](https://nikolasent.github.io/proj/comp2)__.

The solution is based on a LinkNet neural network for semantic segmentation. Loss function is based on a weighted pixel-wise F-eta scores.

The approach contains some inference speed up techniques as FPS was an essential part of the competition.

### Requirements

[PyTorch 0.4](https://pytorch.org/) framework with torchvision was used for the project. Modern Nvidia GPU with CUDA 9.1 with 3 patches installed is required for correct and fast training.

Python 3.6 and several Python modules:

```

pybase64

joblib

opencv-python

numpy

matplotlib

```

### How to train

The whole training pipeline is implemented in [train.py](https://github.com/NikolasEnt/Lyft-Perception-Challenge/blob/master/train.py). Just list train and val folders in the train_dirs and val_dirs lists and run ```python train.py``` to recreate the final submission model. Several sample images are provided with the repo, see [data](https://github.com/NikolasEnt/Lyft-Perception-Challenge/tree/master/data). Each dataset folder should contain "CameraRGB" folder with images and a "CameraSeg" folder with segmentation masks from the simulator.

__Note:__ It may take about two days to train on 15601 images in the train dataset and 1500 images in the val dataset with a single Nvidia GTX 1080 Ti GPU.

For experiments and data visualization, use the [train.ipynb](https://github.com/NikolasEnt/Lyft-Perception-Challenge/blob/master/train.ipynb). It also may be useful to save the hood mask.

### How to predict

A client-server approach was used for the inferencestage, that is why one should start a server with correct [model path](https://github.com/NikolasEnt/Lyft-Perception-Challenge/blob/0c042ed05b11f31304df2e73676f2c0aaaf34382/predict-server.py#L17):

```

python predict-server.py

```

One may need to adjust batch size (`batch`) and number of threads for parallel prediction encoding (`n_jobs`) for performance maximization on a given hardware setup.

And after it is ready, the client can be used:

```

# In another terminal

python predict-client.py path/to/the/test/video

```

It will print out the prediction results as base64 masks encoded with png. To save the predictions as a json file, uncomment [lines 12-13](https://github.com/NikolasEnt/Lyft-Perception-Challenge/blob/0c042ed05b11f31304df2e73676f2c0aaaf34382/predict-client.py#L12) in the predict-client.py.

Current implementation takes 800x600 px video as an input file.