Acoustic-Image-Generation

Code for the paper "Audio-Visual Localization by Acoustic Image Generation", AAAI 2021

Requirements

• Python 3
• TensorFlow 1.14.0 >=

Contents

We provide several scripts and a package with all the necessary code for training and testing our model. The code is organized in several folders and a couple of main scripts as follows:

• The main script is used for training and testing the different models: UNet, DualCamNet with real and generated images.

• The showimages_bb plots FlickrSoundnet energy from a UNet checkpoint and list of tfrecords of FlickrSoundnet.

• The showimages plots ACIVW and AVIA energy from a UNet checkpoint and list of testing tfrecords.

• The showvideo plots VGG Sound energy from a UNet checkpoint and list of tfrecords of a video.

• The meanstd computes the mean the computed metrics of 5 experiments excluding for each metric min and max values and saves them in a xlsx file.