From 23156cf284c27daa825cf987b1ba252edd9a90b7 Mon Sep 17 00:00:00 2001 From: Haojin Yang Date: Wed, 9 Feb 2022 22:45:03 +0100 Subject: [PATCH] improved readme test --- masked_face/README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/masked_face/README.md b/masked_face/README.md index fcf13c9..becbc99 100644 --- a/masked_face/README.md +++ b/masked_face/README.md @@ -16,7 +16,7 @@ However, this model is trained only using 4095 images (2165 masked / 1930 withou We will experience many false positives in the actual application scenarios. This original intention inspired me to build a larger dataset to provide better open-source masked facial detection models and help the world survive the pandemic. -The main contribution of this project is to provide balanced facial training data combining the [Wider face](https://drive.google.com/open?id=1OBY-Pk5hkcVBX1dRBOeLI4e4OCvqJRnH) and [MAFA face](https://imsg.ac.cn/research/maskedface.html) dataset. The [MAFA](https://imsg.ac.cn/research/maskedface.html) data was converted to pascal-VOC format and merged into the [wider_face_add_lm_10_10](https://drive.google.com/open?id=1OBY-Pk5hkcVBX1dRBOeLI4e4OCvqJRnH). +The main contribution of this project is to provide balanced facial training data combining the [wider_face_add_lm_10_10](https://drive.google.com/open?id=1OBY-Pk5hkcVBX1dRBOeLI4e4OCvqJRnH) and [MAFA face](https://imsg.ac.cn/research/maskedface.html) dataset. The [MAFA](https://imsg.ac.cn/research/maskedface.html) data was converted to pascal-VOC format and merged into the [wider_face_add_lm_10_10](https://drive.google.com/open?id=1OBY-Pk5hkcVBX1dRBOeLI4e4OCvqJRnH). ## About the WIDER_MAFA_Balanced dataset The *Wider_MAFA_Balanced* dataset (**4.8GB**) can be downloaded at [HPI owncloud](https://owncloud.hpi.de/s/L4MUGqrpeENLbSv). @@ -28,6 +28,7 @@ Source| Class | Train | Test |Total| MAFA face| masked_face | 15542 | 3922 | 19464 | Wider face| face | 12859 | 3219 | 16078 | *MAFA human body* | face | 2683 | 0 |2683 + *MAFA human body* indicates the extracted training samples with human body occlusions. I use this script for converting MAFA data format to pascal VOC: