previtus / dataset_handlers

small scripts+tricks to assemble+visualize datasets fast

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Dataset tools

Small scripts+tricks to assemble+visualize your datasets fast. We have two versions right now - 1.) making a video and 2.) making a t-SNE grid.

Video preview of the dataset

Installation

This code uses and requires ffmpeg to be installed.

On Mac use: brew install ffmpeg

On Linux use:

sudo add-apt-repository ppa:jonathonf/ffmpeg-4
sudo apt-get update
sudo apt-get install ffmpeg

On Windows: follow a guide (maybe this one)

Test your installation by running this in the terminal: ffmpeg -version

Using the code

Open the file ffmpeg_video.py and change the paths so they point where you have the images:

It is this line:

images_path = "/home/vitek/Projects/dataset_handlers/DATASETS/inputs_v4" << REPLACE THIS WHOLE PATH

PS: The code will look for images (jpg or png) in this directory and all folders under it.

After editing the path run:

python ffmpeg_video.py

If everything went well, you'll have your own version of output_timeOrder.avi as the output

What could go wrong:

If the folder doesn't have any images or the images are not jpg or png. If the ffmpeg is not setup correctly.

About

small scripts+tricks to assemble+visualize datasets fast


Languages

Language:Python 100.0%