Resmasknet

🚀 ResMaskNet

ResMaskNet结合了残差掩码和U-Net架构，可从图像中预测7种面部情绪类别，为面部表情识别提供了有效的解决方案。

🚀 快速开始

以下是使用ResMaskNet进行面部情绪分类的示例代码：

import numpy as np
import torch
import torch.nn as nn
from feat.emo_detectors.ResMaskNet.resmasknet_test import ResMasking
from huggingface_hub import hf_hub_download

# Load Configs
emotion_config_file = hf_hub_download(repo_id= "py-feat/resmasknet", filename="config.json", cache_dir=get_resource_path())
with open(emotion_config_file, "r") as f:
    emotion_config = json.load(f)

device = 'cpu'
emotion_detector = ResMasking("", in_channels=emotion_config['in_channels'])
emotion_detector.fc = nn.Sequential(nn.Dropout(0.4), nn.Linear(512, emotion_config['num_classes']))
emotion_model_file = hf_hub_download(repo_id='py-feat/resmasknet', filename="ResMaskNet_Z_resmasking_dropout1_rot30.pth")
emotion_checkpoint = torch.load(emotion_model_file, map_location=device)["net"]
emotion_detector.load_state_dict(emotion_checkpoint)
emotion_detector.eval()
emotion_detector.to(device)


# Test model
face_image = "path/to/your/test_image.jpg"  # Replace with your extracted face image that is [224, 224]

# Classification - [angry, disgust, fear, happy, sad, surprise, neutral]
emotions = emotion_detector.forward(face_image)
emotion_probabilities = torch.softmax(emotions, 1)
        

✨ 主要特性

• 结合残差掩码和U-Net架构，用于面部情绪分类。
• 能够从图像中准确预测7种面部情绪类别。

📚 详细文档

模型描述

ResMaskNet结合了残差掩码和U-Net架构，可从图像中预测7种面部情绪类别。

模型详情

模型来源

• 仓库：GitHub仓库
• 论文：Facial Expression Recognition Using Residual Masking Network

引用

如果您在研究或应用中使用了svm_au模型，请引用以下论文：

Pham Luan, The Huynh Vu, and Tuan Anh Tran. "Facial Expression Recognition using Residual Masking Network". In: Proc. ICPR. 2020.

@inproceedings{pham2021facial,
  title={Facial expression recognition using residual masking network},
  author={Pham, Luan and Vu, The Huynh and Tran, Tuan Anh},
  booktitle={2020 25th International Conference on Pattern Recognition (ICPR)},
  pages={4513--4519},
  year={2021},
  organization={IEEE}
}

致谢

我们感谢Luan Pham以宽松的许可证慷慨分享此模型。

📄 许可证

本模型使用MIT许可证。