医院手术器械少样本图像检索的可行性研究

      Feasibility of Few-Shot Image Retrieval for Hospital Surgical Instruments

      • 摘要:
        目的 针对医院手术器械文字档案与实物图像缺少关联的问题,构建少样本图像检索辅助流程,评价有限图库下的类别检索可行性。
        方法 使用自建5类487张和公开6类3840张器械图像,以冻结的DFN5B-CLIP为视觉骨干,融合完整原图与锚定区域特征,结合最近实例和类别中心相似度进行排序。在自建开发子集选择融合权重并锁定后,对两套数据分别进行10次嵌套随机图库抽样,比较每类1、3、5张图像下的检索性能。
        结果 每类5张时,自建与公开数据的首位召回率分别为(99.62±0.45)%和(91.69±2.88)%;公开数据的类别级平均精度为(95.45±1.51)%。相同配置下,DFN5B-CLIP的三项检索指标均值高于其余三种骨干。
        结论 该方法可在有限图库下提供候选类别排序,为器械图像查询与人工复核提供技术支持。

         

        Abstract:
        Objective To develop a few-shot image retrieval workflow addressing the lack of links between textual records and physical images of hospital surgical instruments, and to evaluate category retrieval with limited galleries.
        Methods A self-built dataset of 487 images from five categories and a public dataset of 3840 images from six categories were used. A frozen DFN5B-CLIP visual backbone combined features from original images and anchor-region images and ranked categories using nearest-instance and class-prototype similarities. Fusion weights were selected on the self-built development subset and then locked. Ten nested random gallery samplings were conducted on each dataset to compare retrieval with one, three, or five images per category.
        Results With five images per category, Recall@1 was (99.62±0.45)% on the self-built dataset and (91.69±2.88)% on the public dataset; class-level mAP on the public dataset was (95.45±1.51)% (mean±standard deviation). Under the same configuration, DFN5B-CLIP had higher mean values for all three retrieval metrics than the other three backbones.
        Conclusion The method provides candidate category rankings with limited galleries and technical support for instrument image queries and manual review.

         

      /

      返回文章
      返回