TensorFlow SiameseNet Face Recognition

最新推荐文章于 2023-04-14 11:44:32 发布

原创

最新推荐文章于 2023-04-14 11:44:32 发布 · 716 阅读

0 ·

CC 4.0 BY-SA版权

文章标签：

#TensorFlow #SiameseNet #Face #Recognition

本文主要工作是利用SiameseNet模型实现人脸识别功能。采用简化版VGG的SiameseNet卷积神经网络模型，数据集使用微软的MS-Celeb-1M公开人脸数据集，训练在TITAN X上约3天，损失函数为Logistic Regression，最终在LFW上的准确率超过90%，但模型仍有优化空间。

The main work of the article is to use the SiameseNet model to achieve the function of face recognition.
The SiameseNet convolutional neural network model is as follows (detailed structure moves to GitHub, a simplified version of VGG):

Input and output examples

The data set uses Microsoft’s MS-Celeb-1M public face data set. We directly download the aligned data set. First, there is a cleaned list on the network. The file name is: MS-Celeb-1M_clean_list.txt, about 160M. Then use the combination algorithm to generate two files, positive_pairs_path.txt and negative_pairs_path.txt, each about 1.5million pairs.
Generate file example