DDQA Dataset (Defect Detection Question-Answering dataset)
收藏资源简介:
DDQA数据集是由上海交通大学团队构建的多模态工业异常检测训练数据集,包含127997张图像,涵盖多种对象类别和复杂场景。该数据集根据现有数据集中的信息,按照特定规则构建,不使用生成模型,有效减少了数据噪声,提高了数据的准确性和可信度。数据集包含四种任务:异常识别、粗糙缺陷定位、缺陷精细映射和缺陷分类,以增强模型在工业质量检测方面的能力。
The DDQA dataset is a multimodal industrial anomaly detection training dataset constructed by the team from Shanghai Jiao Tong University. It contains 127,997 images covering multiple object categories and complex scenarios. This dataset is built based on information from existing datasets following specific rules, without using generative models, which effectively reduces data noise and improves the accuracy and credibility of the dataset. The dataset includes four tasks: anomaly recognition, coarse defect localization, fine defect mapping, and defect classification, aiming to enhance the model's capabilities in industrial quality inspection.




