Visual question answering: Datasets, algorithms, and future challenges

Kushal Kafle, Christopher Kanan

Abstract: Highlights • Comparison of visual question answering (VQA) with related computer vision tasks. • Critical review of all major VQA datasets and evaluation metrics. • Comprehensive review and comparison of existing methods for VQA. • All major datasets have language and difficulty bias that critically affects VQA. • Recommendations for future VQA datasets and evaluation metrics to combat bias. Abstract Visual Question Answering (VQA) is a recent problem in computer vision and natural language processing that has garnered a large amount of interest from the deep learning, computer vision, and natural language processing communities. In VQA, an algorithm needs to answer text-based questions about images. Since the release of the first VQA dataset in 2014, additional datasets have been released and many algorithms have been proposed. In this review, we critically examine the current state of VQA in terms of problem formulation, existing datasets, evaluation metrics, and algorithms. In particular, we discuss the limitations of current datasets with regard to their ability to properly train and assess VQA algorithms. We then exhaustively review existing algorithms for VQA. Finally, we discuss possible future directions for VQA and image understanding research.