通过面部重组进行年龄和性别检测的内部工具。 [英] In house tool for age and gender detection through face reorganization.

问题描述

大家好，

希望你们都做得很好，我们正在使用wpf开发一个标牌应用程序。根据客户的要求，我们希望检测正在查看在我们的标牌上运行的广告的人的年龄和性别。但问题是由于一些法律问题我们不应该将这些图像发送到第三方云服务器进行处理（通过api）。所以任何人都可以建议我最好的内部工具或sdk（付费/开源）作为解决方案。

谢谢。



我尝试过：



---------------------- -------------------------------------------------- ----------------

Hi Guys,
Hope u all doing well, We guys are developing a signage application using wpf. As per the requirement form the client we want to detect the age and gender of the people who are looking at the advertisements running on our signage. But the problem is due to some legal issue we are not supposed to send these image to a third party cloud server for processing(through api's). So can any one suggest me a best in house house tool or sdk (Paid /Open Source) as a solution.
Thanks.

What I have tried:

----------------------------------------------------------------------------------------

推荐答案

为什么不直接存储要评估的图像部分而不是整个画面。



- 或 -



将每张照片分成足够数量的EQUAL-SIZE列通过仅查看一列来使主题无法识别，并分别存储每一列，将它们拼接在一起以便稍后进行评估。



UPDATE ===== ==================



既然你似乎不理解......



将图像分成几部分：



例如，如果您的图像宽度为600像素，则可能会破坏图像到任意数量的像素均匀地分成600.例如，2,4,5,6,10,15,20,25,30,40,50,60,120等）。



存储云中的部分：



将所有这些图像部分发送到云端进行存储。个别地，他们没有呈现被​​评估的人的形象，所以你没有违反任何法律。如果需要更高的安全性，您可以选择加密每个部分。



从云中检索部分：



检索部分并重新组装它们。如果您正确地创建云存储架构，这应该是一个小练习。



评估图像：



做任何魔法来确定性别/年龄，然后丢弃重新组装的图像，然后从云端删除部件。



---- --------



您将不得不实际编写代码将图像分成几部分，然后重新组装它们，但是真的不应该那么难。有关如何在互联网上操纵图像的大量信息。谷歌是免费的，所以请考虑这是一个为你自我发现的新时代的邀请。

Why don't you just store the parts of the image that you will be evaluating instead of the whole picture.

- or -

Break each photo up into a sufficient number of EQUAL-SIZE columns to make the subject unrecognizable by viewing just one column, and store each column separately, stitching them together for evaluation later.

UPDATE =======================

Since you didn't seem to understand...

Break the image into parts:

As an example, if your image is 600 pixels wide, you could break the image up into any number of pixels that is evenly divisible into 600. For instance 2, 4, 5, 6, 10, 15, 20, 25, 30, 40, 50, 60, 120, etc).

Store the Sections In the Cloud:

Send all of these image sections to the cloud for storage. Individually, they do not present an image of the person being "evaluated", so you're not breaking any laws. You could optionally encrypt each section if more security is desired.

Retrieve the Sections from the Cloud:

Retrieve the sections and re-assemble them. This should be a minor exercise if you create your cloud storage schema appropriately.

Evaluate the Image:

Do whatever magic you do to determine gender/age, and then dispose the re-assembled image, and then delete the parts from the cloud.

------------

You're going to have to actually write code to break the image into parts, and then re-assemble them, but that really should not be that difficult. There's plenty of info on how to manipulate images on the internet. Google is free, so consider this an invitation to a new era of self-discovery for you.

这篇关于通过面部重组进行年龄和性别检测的内部工具。的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！