CT-RATE-VQA
收藏资源简介:
CT-RATE-VQA数据集是一个基于CT图像的视觉问答数据集,包含84,500个问答对。该数据集旨在支持医学影像模型训练和评估,涵盖了七种临床常见疾病类别,如气道异常、肺气肿、纤维化相关改变等。数据集通过局部放大技术增强了病变特征,并利用强化学习框架提高了模型的推理能力,从而实现了对胸片异常的更准确和稳定的识别。
The CT-RATE-VQA dataset is a computed tomography (CT) image-based visual question answering (VQA) dataset containing 84,500 question-answer pairs. It is designed to support the training and evaluation of medical imaging models, covering seven clinically common disease categories such as airway abnormalities, emphysema, fibrosis-related alterations, and others. The dataset enhances lesion features through local magnification techniques, and leverages a reinforcement learning framework to improve the reasoning capabilities of models, thereby enabling more accurate and stable recognition of chest X-ray abnormalities.

- 1通过中国科学院 · 2025年



