按内容比较大量图像 [英] Comparing a large set of images by content

问题描述

我正在使用Flask（Python）运行服务器，允许用户上传图像。 （最终图像将被发送到S3存储桶;现在，它们只是保存到磁盘。）但是，我想防止重复和类似的图像。

I am running a server using Flask (Python) that allows users to upload images. (Eventually images will be sent to an S3 bucket; for now, they're just saved to disk.) However, I want to prevent duplicate and similar images.

这意味着每次用户尝试上传图片时，我都要搜索数据库（类似于 tineye ）。如果存在类似的结果，则应用程序将不会上传图像。只有Python Imaging Library才能实现这种反向图像搜索吗？是否有其他用于完成类似任务的Python框架？或者也许是亚马逊工具？

This means that each time that a user attempts to upload an image, I have to search through the database (in a fashion similar to tineye). If a similar result exists, the app will not upload the image. Is this type of "reverse image search" possible with just the Python Imaging Library? Are there other Python frameworks out there that have been used to accomplish similar tasks? Or perhaps an Amazon tool?

推荐答案

使用图像库创建独特图像的指纹。如果要检查新图像针对指纹并适当地做出决定。

Using the image library create a finger print of unique images.If a new image comes in check against the finger prints and take the decision appropriately.

对于指纹识别，请参考此
https://realpython.com/blog/python/fingerprinting-images-for-near-duplicate-detection/

For fingerprinting refer this https://realpython.com/blog/python/fingerprinting-images-for-near-duplicate-detection/

如果你想把AI带到这里，
注意自动编码器。

If you want to bring AI here, Look out for autoencoder.

这篇关于按内容比较大量图像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！