AIC (AI Challenger)
收藏资源简介:
AIC (AI Challenger) 是由创新工场等机构创建的大型图像理解数据集,包含三个子数据集:人体关键点检测(HKD)、大规模属性数据集(LAD)和图像中文描述(ICC)。该数据集通过丰富的标注,如类别标签、关键点坐标、边界框、属性和描述,有效地连接了图像的低级特征与高级概念。数据集总计包含681658张图像,其中HKD包含300,000张,LAD包含81,658张,ICC包含300,000张。这些数据集不仅为评估和改进计算方法提供了有效的基准,也为相关任务提供了新的资源,特别是在人体关键点检测、属性基零样本识别和图像中文描述等领域。
AIC (AI Challenger) is a large-scale image understanding dataset created by institutions including Innovation Works. It encompasses three sub-datasets: Human Keypoint Detection (HKD), Large-scale Attribute Dataset (LAD), and Image Chinese Captioning (ICC). Rich annotations such as category labels, keypoint coordinates, bounding boxes, attributes and textual descriptions enable the dataset to effectively bridge the low-level visual features and high-level semantic concepts of images. In total, the dataset contains 681,658 images, with 300,000 in HKD, 81,658 in LAD, and 300,000 in ICC. These datasets not only provide valid benchmarks for evaluating and refining computational approaches, but also offer new resources for related tasks, especially in the fields of human keypoint detection, attribute-based zero-shot recognition, and image Chinese captioning.




