Learning from Multiple Expert Annotators for Enhancing Anomaly Detection in Medical Image Analysis

doi:doi:10.57702/4gst46z5

Learning from Multiple Expert Annotators for Enhancing Anomaly Detection in Medical Image Analysis

The proposed method uses Weighted Boxes Fusion (WBF) algorithm to obtain the aggregated annotations with the implicit annotators' agreement as conﬁdence scores. The estimated annotations are then used to train a deep learning detector with a re-weighted loss function that incorporates the conﬁdence scores to localize abnormal ﬁndings.

BibTex:

@dataset{Khiem_H_Le_and_Tuan_V_Tran_and_Hieu_H_Pham_and_Hieu_T_Nguyen_and_Tung_T_Le_and_Ha_Q_Nguyen_and_D_Q_Tran_and_D_B_Nguyen_and_C_M_Pham_and_H_T_Tong_and_D_H_Dinh_2025,
    abstract = {The proposed method uses Weighted Boxes Fusion (WBF) algorithm to obtain the aggregated annotations with the implicit annotators' agreement as conﬁdence scores. The estimated annotations are then used to train a deep learning detector with a re-weighted loss function that incorporates the conﬁdence scores to localize abnormal ﬁndings.},
    author = {Khiem H. Le and Tuan V. Tran and Hieu H. Pham and Hieu T. Nguyen and Tung T. Le and Ha Q. Nguyen and D. Q. Tran and D. B. Nguyen and C. M. Pham and H. T. Tong and D. H. Dinh},
    doi = {10.57702/4gst46z5},
    institution = {No Organization},
    keyword = {'Anomaly Detection', 'Deep Learning', 'Medical Image Analysis', 'Object Detection'},
    month = {jan},
    publisher = {TIB},
    title = {Learning from Multiple Expert Annotators for Enhancing Anomaly Detection in Medical Image Analysis},
    url = {https://service.tib.eu/ldmservice/dataset/learning-from-multiple-expert-annotators-for-enhancing-anomaly-detection-in-medical-image-analysis},
    year = {2025}
}