Experiment related to research scholarship.
The How2 corpus (SANABRIA et al., 2018) was used to validate the model proposed in my research project, so it was necessary to perform some pre-processing steps, such as extracting image frames and their respective bilingual captions, bilingual textual alignment and multimodal text-image alignment.
This repository contains some files generated during the experiment.