X-Ray Image Contraband Detection Dataset at Security Checkpoints

This dataset boasts extremely high data quality, with annotation accuracy exceeding 98%, ensuring the consistency and completeness of labeled data. Technically, it introduces advanced multi-view annotation and data augmentation techniques for the first time, enhancing the model's generalization ability on small samples. In practical applications, models trained on this dataset can effectively reduce miss and false detection rates, improving security check efficiency. Moreover, compared to other similar datasets, this dataset has significant advantages in terms of data variety and complexity, covering a wider range of contraband types and security check scenarios. Additionally, the dataset contains a large number of samples of rare items, making its data characteristics unique and irreplaceable. Due to its reasonable structured data design, it has high scalability and versatility, making it easy to integrate and apply in future systems.