Ziran Zhang^1,2, Yongrui Ma^3,2, Yueting Chen¹, Feng Zhang², Jinwei Gu³, Tianfan Xue³, Shi Guo²

¹ Zhejiang University, ² Shanghai AI Laboratory, ³ The Chinese University of Hong Kong

🌐 Project Page | 🎥 Video | 📄 Paper | 📊 Data | 🛠️ Weights | 🔖 PPT

🔔 Overview

This repository hosts the implementation of "From Sim-to-Real: Toward General Event-based Low-Light Frame Interpolation with Per-scene Optimization" (SIGGRAPH Asia 2024). Our approach leverages event cameras and enhances Video Frame Interpolation (VFI) in low-light conditions via a per-scene optimization strategy. This method adapts the model to specific lighting and camera settings, solving issues like trailing artifacts and signal degradation common in low-light environments.

🔑 Key Features

• Per-Scene Optimization: Fine-tunes a pre-trained model for each scene, significantly improving interpolation results in varied lighting conditions.
• Low-Light Event Correction: Effectively mitigates event-based signal latency and noise under low-light conditions.
• EVFI-LL Dataset: Provides challenging RGB+Event sequences captured in low-light environments for benchmarking.

🚀 Quick Start: Per-Scene Optimization

Follow these steps to apply per-scene optimization with pre-trained models:

1. Clone the repository:

git clone https://github.com/OpenImagingLab/Sim2Real.git
cd Sim2Real/Sim2Real_code

2. Download and place the required weights:

• Main model weights: Place in pretrained_weights/
• DISTS loss function weights: Place in losses/DISTS/weights/

Download Weights Here

3. Install dependencies:

Ensure Python 3.9.19 is installed, then run:

pip install -r requirements.txt

4. Run per-scene optimization

To run the per-scene optimization, follow these steps:

Modify the paths in the configuration file:

Open the Sim2Real_code/params/Paths/RealCaptured.py file and update the paths to point to the actual location where you have downloaded your dataset. Modify the following lines:

RC.train.rgb = "/ailab/user/zhangziran/Dataset/Sim2Real_release"
RC.train.evs = "/ailab/user/zhangziran/Dataset/Sim2Real_release"
RC.test.rgb = "/ailab/user/zhangziran/Dataset/Sim2Real_release"
RC.test.evs = "/ailab/user/zhangziran/Dataset/Sim2Real_release"

Change them to the correct path on your system.

Run the optimization script:

After modifying the paths, execute the following command to start the per-scene optimization process:

bash perscene.sh

This will fine-tune the pre-trained model on specific scenes, performing frame interpolation optimized for each setting.

🛠️ Optional: Pretraining the Model

To pretrain the model from scratch using simulated data:

1. Pretrain the model:

bash pretrain.sh

2. After pretraining, proceed with per-scene optimization as described above.

📂 Directory Structure

• dataset/: Utilities for dataset preparation and loading.
• losses/: Custom loss functions and weights for training.
• models/: Neural network models for Sim2Real frame interpolation tasks.
• params/: Configuration files for training and evaluation.
• tools/: Scripts for preprocessing and postprocessing.
• pretrained_weights/: Directory for storing pre-trained models.
• run_network.py: Main script for training and evaluation.
• pretrain.sh: Script for model pretraining.
• perscene.sh: Script for per-scene optimization.
• requirements.txt: Required Python dependencies.

📂 Simulated Data Preparation

This section outlines the steps for preparing the simulated data:

1. Frame Interpolation with RIFE
   First, use RIFE to perform 8x frame interpolation on the GoPro dataset.

2. Converting RGB Frames to RAW Domain
   Next, use inv_isp to convert the dense RGB frames to the RAW domain.

   • Modify the inv_isp.py script.
   • Set the appropriate paths in the script.
   • Run the script to process the data.

3. Generating Simulated Event Signals
   Finally, use the v2e_idslpf simulator to generate the simulated event signals.

   • Modify the v2e_idslpf/config/GOPRO_config.py configuration file.
   • Set the correct paths for the dataset and parameters.
   • Run the simulator to generate the event data.

📊 EVFI-LL Dataset

The EVFI-LL dataset includes RGB+Event sequences captured under low-light conditions, offering a challenging benchmark for evaluating event-based VFI performance. Download and place the dataset in the dataset/ directory.

Download Dataset Here

How to Run a Demo Case

1. Download demo.zip from this link.

2. Extract demo.zip into the data folder.

3. Update the path in Sim2Real/params/Paths/RealCaptured.py to point to the /data folder.

4. In Sim2Real/Sim2Real_code/dataset/RC/dataset_dict.py, keep only "demo" in both dataset_dict and test_key.

5. Please modify the save path in Sim2Real/Sim2Real_code/params/RealCaptured.py, and update the number of test cases in perscene.sh by setting:

for idx in {0..0}

6. Run the code.

📜 License Information

The code in this repository is licensed under the MIT License.

📝 Citation

If you find this work helpful in your research, please cite:

@article{zhang2024sim,
  title={From Sim-to-Real: Toward General Event-based Low-light Frame Interpolation with Per-scene Optimization},
  author={Zhang, Ziran and Ma, Yongrui and Chen, Yueting and Zhang, Feng and Gu, Jinwei and Xue, Tianfan and Guo, Shi},
  journal={arXiv preprint arXiv:2406.08090},
  year={2024}
}

🙏 Acknowledgements

This project builds upon the exceptional work of TimeLens-XL. We extend our sincere thanks to the original authors for their outstanding contributions.