Robust Video Matting (RVM)
English | 中文
Official repository for the paper Robust High-Resolution Video Matting with Temporal Guidance. RVM is specifically designed for robust human video matting. Unlike existing neural models that process frames as independent images, RVM uses a recurrent neural network to process videos with temporal memory. RVM can perform matting in real-time on any videos without additional inputs. It achieves 4K 76FPS and HD 104FPS on an Nvidia GTX 1080 Ti GPU. The project was developed at ByteDance Inc.
News
- [Aug 25 2021] Source code and pretrained models are published.
- [Jul 27 2021] Paper is accepted by WACV 2022.
Showreel
Watch the showreel video (YouTube, Bilibili) to see the model's performance.
All footage in the video are available in Google Drive and Baidu Pan (code: tb3w).
Demo
- Webcam Demo: Run the model live in your browser. Visualize recurrent states.
- Colab Demo: Test our model on your own videos with free GPU.
Download
We recommend MobileNetv3 models for most use cases. ResNet50 models are the larger variant with small performance improvements. Our model is available on various inference frameworks. See inference documentation for more instructions.
| Framework | Download | Notes |
| PyTorch | rvm_mobilenetv3.pth rvm_resnet50.pth |
Official weights for PyTorch. Doc |
| TorchHub | Nothing to Download. | Easiest way to use our model in your PyTorch project. Doc |
| TorchScript | rvm_mobilenetv3_fp32.torchscript rvm_mobilenetv3_fp16.torchscript rvm_resnet50_fp32.torchscript rvm_resnet50_fp16.torchscript |
If inference on mobile, consider export int8 quantized models yourself. Doc |
| ONNX | rvm_mobilenetv3_fp32.onnx rvm_mobilenetv3_fp16.onnx rvm_resnet50_fp32.onnx rvm_resnet50_fp16.onnx |
Tested on ONNX Runtime with CPU and CUDA backends. Provided models use opset 12. Doc, Exporter. |
| TensorFlow | rvm_mobilenetv3_tf.zip rvm_resnet50_tf.zip |
TensorFlow 2 SavedModel. Doc |
| TensorFlow.js | rvm_mobilenetv3_tfjs_int8.zip |
Run the model on the web. Demo, Starter Code |
| CoreML | rvm_mobilenetv3_1280x720_s0.375_fp16.mlmodel rvm_mobilenetv3_1280x720_s0.375_int8.mlmodel rvm_mobilenetv3_1920x1080_s0.25_fp16.mlmodel rvm_mobilenetv3_1920x1080_s0.25_int8.mlmodel |
CoreML does not support dynamic resolution. Other resolutions can be exported yourself. Models require iOS 13+. s denotes downsample_ratio. Doc, Exporter |
All models are available in Google Drive and Baidu Pan (code: gym7).
PyTorch Example
- Install dependencies:
pip install -r requirements_inference.txt
- Load the model:
import torch
from model import MattingNetwork
model = MattingNetwork('mobilenetv3').eval().cuda() # or "resnet50"
model.load_state_dict(torch.load('rvm_mobilenetv3.pth'))
- To convert videos, we provide a simple conversion API:
from inference import convert_video
convert_video(
model, # The model, can be on any device (cpu or cuda).
input_source='input.mp4', # A video file or an image sequence directory.
output_type='video', # Choose "video" or "png_sequence"
output_composition='output.mp4', # File path if video; directory path if png sequence.
output_video_mbps=4, # Output video mbps. Not needed for png sequence.
downsample_ratio=None, # A hyperparameter to adjust or use None for auto.
seq_chunk=12, # Process n frames at once for better parallelism.
)
- Or write your own inference code:
from torch.utils.data import DataLoader
from torchvision.transforms import ToTensor
from inference_utils import VideoReader, VideoWriter
reader = VideoReader('input.mp4', transform=ToTensor())
writer = VideoWriter('output.mp4', frame_rate=30)
bgr = torch.tensor([.47, 1, .6]).view(3, 1, 1).cuda() # Green background.
rec = [None] * 4 # Initial recurrent states.
downsample_ratio = 0.25 # Adjust based on your video.
with torch.no_grad():
for src in DataLoader(reader): # RGB tensor normalized to 0 ~ 1.
fgr, pha, *rec = model(src.cuda(), *rec, downsample_ratio) # Cycle the recurrent states.
com = fgr * pha + bgr * (1 - pha) # Composite to green background.
writer.write(com) # Write frame.
- The models and converter API are also available through TorchHub.
# Load the model.
model = torch.hub.load("PeterL1n/RobustVideoMatting", "mobilenetv3") # or "resnet50"
# Converter API.
convert_video = torch.hub.load("PeterL1n/RobustVideoMatting", "converter")
Please see inference documentation for details on downsample_ratio hyperparameter, more converter arguments, and more advanced usage.
Training and Evaluation
Please refer to the training documentation to train and evaluate your own model.
Speed
Speed is measured with inference_speed_test.py for reference.
| GPU | dType | HD (1920x1080) | 4K (3840x2160) |
|---|---|---|---|
| RTX 3090 | FP16 | 172 FPS | 154 FPS |
| RTX 2060 Super | FP16 | 134 FPS | 108 FPS |
| GTX 1080 Ti | FP32 | 104 FPS | 74 FPS |
- Note 1: HD uses
downsample_ratio=0.25, 4K usesdownsample_ratio=0.125. All tests use batch size 1 and frame chunk 1. - Note 2: GPUs before Turing architecture does not support FP16 inference, so GTX 1080 Ti uses FP32.
- Note 3: We only measure tensor throughput. The provided video conversion script in this repo is expected to be much slower, because it does not utilize hardware video encoding/decoding and does not have the tensor transfer done on parallel threads. If you are interested in implementing hardware video encoding/decoding in Python, please refer to PyNvCodec.
10 Aug 30, 2022
8 Aug 18, 2022
71 Dec 19, 2022
3 Jan 11, 2022
10 Jul 16, 2022
4 Feb 22, 2022
133 Jan 05, 2023
217 Dec 05, 2022
22 Dec 13, 2022
543 Dec 30, 2022
4 Aug 30, 2022
18 Nov 06, 2022
101 Dec 27, 2022
17 Dec 29, 2022
2 May 05, 2022
3.2k Dec 31, 2022
0 Jul 15, 2021
45 Dec 28, 2022
2 Apr 23, 2022
8 Dec 13, 2022
23 Oct 08, 2022
32 Oct 24, 2022
13 Dec 25, 2022
7 Dec 12, 2022