vision-rekognition

This repository lets you compare image content analysis performed by Google Vísion and AWS Rekognition

How use it?

Configurations

Clone this project in your machine

Rekognition

Get your credentials in AWS console and create the config.yml file in root path

config.yml

AWS_REGION: us-east-1
AWS_ACCESS_KEY_ID: AAAAAAAAAAA
AWS_SECRET_ACCESS_KEY: AAAAAAAAAA

Vision

Set up a project in Google Cloud and enable the Vision API, after obtaining a credential file, save it to the root path with name vision-credentials.json

Images

Add sample images for comparison in image folder

Exec

At this point you should have something like this in the project

images/
  01.jpg
  02.jpg
config.yml
vision-credentials.json

So, execute make run

$ make run
go build -o vrekog . && ./vrekog
02.jpg -- Vision: Medical: 2.00 Racy: 5.00 Spoof: 1.00 Violence: 1.00 Adult: 3.00
02.jpg -- REKOG:
------------------------
01.jpg -- Vision: Adult: 2.00 Medical: 1.00 Racy: 4.00 Spoof: 1.00 Violence: 1.00
01.jpg -- REKOG: Revealing Clothes: 75.38 Suggestive: 75.38
------------------------
03.jpg -- REKOG: Graphic Female Nudity: 99.88 Explicit Nudity: 99.88 Sexual Activity: 94.61
03.jpg -- Vision: Adult: 5.00 Medical: 1.00 Racy: 5.00 Spoof: 1.00 Violence: 1.00