当前位置:网站首页>Google is improving the skin color performance in all products and practicing the concept of "image fairness"
Google is improving the skin color performance in all products and practicing the concept of "image fairness"
2022-07-23 18:49:00 【zebra】

today ,Google Decide on its impact on “ Image fairness ” Introduce a new step into the commitment of , And improve the representativeness of all its products .Google Has worked with Ellis, a professional and sociologist at Harvard University - With Dr. monk , The company is releasing a new skin color scale , It aims to be more inclusive of all kinds of skin colors we see in our daily life .

Munk skin color table will completely change the expression of different skin colors , This is due to Google, It is designed to have an easy-to-use technology for development and evaluation ,Google Call it the Munk screen tone scale , You can take a look at it below :

Updating our treatment of skin color can help us better understand the performance in the image , And evaluate whether a product or feature works well in a range of skin colors . This is particularly important for computer vision , This is an artificial intelligence that allows computers to see and understand images . If not intentionally established and tested to include a wide range of skin colors , The computer vision system will be found to be not good enough for people with dark skin .

Google Express , Quantifying skin color will help us and the entire technology industry to build a more representative data set , So that we can train and evaluate the fairness of artificial intelligence model , So as to produce functions and products that are more effective for everyone -- All kinds of skin color . for example , We use this scale to evaluate and improve the face model in the detection image .
You can read more here :
https://blog.google/products/search/monk-skin-tone-scale/
边栏推荐
- 建模刚开始学习很迷茫,次世代角色建模该怎么学习?
- Google正在改进所有产品中的肤色表现 践行“图像公平”理念
- Can self-study 3D modeling succeed? Can self-study lead to employment?
- [attack and defense world web] difficulty Samsung 9-point introductory question (end): Fakebook, favorite_ number
- 【攻防世界WEB】难度三星9分入门题(终):fakebook、favorite_number
- JS convert pseudo array to array
- 【游戏建模模型制作全流程】ZBrush武器模型制作:弩
- PCL:多直線擬合(RANSAC)
- 【游戏建模模型制作全流程】3ds Max和ZBrush制作无线电接收器
- 并非原创的原文路径【如有侵权 请原博主联系删除】
猜你喜欢

【2013】【论文笔记】太赫兹波段纳米颗粒表面增强拉曼——
![MQ [messagequeue graphic explanation and four MQ comparisons]](/img/e0/a8fd8cf3420426c7b1595da79d1b4b.png)
MQ [messagequeue graphic explanation and four MQ comparisons]

建模刚开始学习很迷茫,次世代角色建模该怎么学习?
![Redis [super superfine introductory tutorial]](/img/a6/02b3e670067292d884bab7425b4ffb.png)
Redis [super superfine introductory tutorial]
![Log framework [detailed learning]](/img/2f/2aba5d48e8a544eae0df763d458e84.png)
Log framework [detailed learning]

世界上最大的开源基金会 Apache 是如何运作的?

Three things programmers want to do most | comics

What problems do you usually encounter when learning 3D modeling? It's too much

How does Apache, the world's largest open source foundation, work?

SQLZOO——BBC QUIZ
随机推荐
Three things programmers want to do most | comics
Problems and methods of creating multiple projects under one solution in VS2010
[sharing game modeling model making skills] how ZBrush adjusts the brush size
Detailed explanation of common curl commands and parameters
[the whole process of Game Modeling and model production] create the game soldier character with ZBrush
【2013】【论文笔记】太赫兹波段纳米颗粒表面增强拉曼——
Can self-study 3D modeling succeed? Can self-study lead to employment?
【游戏建模模型制作全流程】3ds Max和ZBrush制作无线电接收器
JUC并发编程【详解及演示】
Analysis on the implementation of Flink exactly once delivery
Rhcsa notes 3
Past weaving of zero one-2022
【2022】【论文笔记】太赫兹量子阱——
Build a PHP development environment (apache+php+mysql) "suggestions collection"
Redis [2022 latest interview question]
学次世代建模是场景好还是角色好?选对职业薪资多一半
1259. 不相交的握手 動態規劃
一文了解 NebulaGraph 上的 Spark 项目
Multithreading [comprehensive study of graphics and text]
【游戏建模模型制作全流程】ZBrush武器模型制作:弩