Recent Advances in Video Question Answering: A Review of Datasets and Methods

Abstract

Video Question Answering (VQA) is a recent emerging challenging task in the field of Computer Vision. Several visual information retrieval techniques like Video Captioning/Description and Video-guided Machine Translation have preceded the task of VQA. VQA helps to retrieve temporal and spatial information from the video scenes and interpret it. In this survey, we review a number of methods and datasets for the task of VQA. To the best of our knowledge, no previous survey has been conducted for the VQA task.

Video Language Co-Attention with Multimodal Fast-Learning Feature Fusion for VideoQA

“…Neural attention mechanisms have become the de-facto standard in machine comprehension tasks (Sood et al, 2020;Yu et al, 2019;Li et al, 2019). In VideoQA, attention mechanisms are particularly important given that the information necessary to generate correct answers is scattered across frames -many of which are redundant or even irrelevant to the question at hand (Patel et al, 2021).…”

“…1 Our code is publicly available at the project website https://www.perceptualui.org/publications/ abdessaied22_repl4NLP/ tried to close by leveraging external memory (Kim et al, 2018(Kim et al, , 2019Fan et al, 2019). While external memory allows models to cache sequential information and retrieve relevant multimodal content (Patel et al, 2021), latest models still suffer from decreased performance, for example on ambiguous questions that require deeper reasoning abilities.…”