Visual Grounding (VG) methods in Visual Question Answering (VQA) attempt to improve VQA performance by strengthening a model's reliance on question-relevant visual information. The presence of such relevant information in the visual input is typically assumed in training and testing. This assumption, however, is inherently flawed when dealing with imperfect image representations common in large-scale VQA, where the information carried by visual features frequently deviates from expected ground-truth contents. As a result, training and testing of VG-methods is performed with largely inaccurate data, which obstructs proper assessment of their potential benefits. In this work, we demonstrate that current evaluation schemes for VG-methods are problematic due to the flawed assumption of availability of relevant visual information. Our experiments show that the potential benefits of these methods are severely underestimated as a result.

翻译：视觉问答（VQA）中的视觉定位（VG）方法旨在通过增强模型对问题相关视觉信息的依赖来提升VQA性能。在训练和测试中，通常假设视觉输入中存在此类相关信息。然而，当处理大规模VQA中常见的非完美图像表示时，这一假设存在根本性缺陷——视觉特征所携带的信息往往偏离预期的真实标注内容。因此，VG方法的训练与测试会基于大量不准确的数据进行，这阻碍了对其潜在优势的合理评估。本研究表明，由于“相关视觉信息可用性”这一假设存在缺陷，当前VG方法的评价体系存在问题。我们的实验证明，这些方法的潜在优势因此被严重低估。