VideoGameQA-Bench

该数据集是一个全面的基准测试，涵盖了广泛的游戏质量保证活动，包括视觉单元测试、视觉回归测试、大海捞针式的任务、故障检测以及为各种游戏图像和视频生成错误报告。此外，该数据集还评估了视觉-语言模型（VLMs）在处理游戏开发中的真实世界场景时的性能表现。其规模广泛，覆盖了多种游戏质量保证活动，任务重点在于视频游戏的质量保证。

This dataset is a comprehensive benchmark covering a wide range of game quality assurance activities, including visual unit testing, visual regression testing, needle-in-a-haystack tasks, fault detection, and generating error reports for various game images and videos. Additionally, this dataset evaluates the performance of Vision-Language Models (VLMs) when handling real-world scenarios in game development. It features extensive coverage across diverse game quality assurance tasks, with the core focus on video game quality assurance.