search

JasonZ95 / image-recognition-using-tweeter-and-google-API

This is a python programme that download specific user's tweeter images and conver it to a video with information attached.
0 stars 0 forks source link

readme

image-recognition-using-tweeter-and-google-API

Instruction

This is a python programme that download user's tweeter images and convert it to a video. On each of the pictures, there are several tags to describe the object on it. To realize the function described above, I apply two APIs:

Tweepy: The API provided by tweeter official to access any user's tweet messages, which is used to download the images for our programme.

Google Vision: An API with powerful image recognition capabilities provided by Google. It can output tags to describe the object in the pictures.

FFMPEG: To convert images into a video in mp4 format.

I also use PIL image library to attach the text message to the pictures, which make the output more friendly.

Preparation

There are several things you need to do before using the programme.

1.Apply for a tweeter developer account to get the key for the tweepy API. The website is 

 https://tweepy.readthedocs.io/en/v3.5.0/

2.Get access for the google cloud platform to get the authentication file (.json) to get access for the API.

3.install relavent resourses(tweepy & PIL), using the following command:

  pip install tweepy

  pip install Pillow

4.Before running the programme, you need to execute following command:

  export GOOGLE_APPLICATION_CREDENTIALS="/directory/googlekey.json"

(This command allows you to use the google vision API.)

  1. After running the programme, you can find some pics in the file you specified. Now you can use following command to convert images to video:

    ffmpeg -f image2 -r 2 -i /directory/image%04d.jpg -vcodec libx264 -t 40 test.mp4