Archives

  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • samples in a each single fold than

    2020-08-18

    samples in a each single fold than that of ten folds. The testing 728
    699 rules is employed to reduce FPs/I to some extent as set is selected in such a way that the proportion of actual 729
    700 demonstrated in Figs. 2 and 3. The results of subjective benign masses to malignant tumours (1:4) is maintained 730
    701 evaluation of the automatic segmentation for detecting approximately. The experiment is performed five times. The 731
    702 suspicious lesions on images of TMC dataset by the expert results are depicted in Table 3. 732
    703 radiologists concerning the biopsy-proven HPRs are enlist- The image-based FPs/I in the single view diagnosis is 1.08 733
    705 As shown in Table 1, the segmentation stage yields total DDSM dataset and results are stored. 735
    708 on MLO view. Thus, the single view CAD algorithm has the
    709 image-based detection sensitivity of (241/263) 91.63% at the The radiologists during their clinical practice observe carefully 737 710 cost of (297/220) 1.35 FPs/I of TMC dataset of 220 images. a single view, either CC or MLO, to find out the abnormal 738 711 Similarly, the results of subjective evaluation of the automatic lesions. If they Erlotinib find the abnormal lesions on one view, they also 739
    Table 3 – Performance of SVM classifier in single view scheme.
    Trial True malignant False malignant True benign False benign
    Please cite this article in press as: Sapate S, et al. Breast cancer diagnosis using abnormalities on ipsilateral views of digital mammograms. Biocybern Biomed Eng (2019), https://doi.org/10.1016/j.bbe.2019.04.008
    Table 4 – Assessment of pairing between CC and MLO view lesions on TMC dataset.
    Correspondence Lesions on Lesions on Possible Pairs within TP–TP pairs with TP–FP pairs FPs/I
    from CC view MLO view pairs annular region highest score
    Table 5 – Assessment of pairing between CC and MLO view lesions on DDSM dataset.
    Correspondence Lesions on Lesions on Possible Pairs within TP–TP pairs with TP–FP pairs FPs/I
    from CC view MLO Erlotinib view pairs annular region highest score
    Table 6 – Performance of SVM classifier using fusion scheme on TMC dataset.
    True malignant False malignant True benign False benign Total pairs Sensitivity FPs/I Kappa
    740 carefully observe the second view. If the abnormality present for training and remaining set of 112 pairs for five-fold cross 777 741 on both the views is likely to represent the same physical validation using SVM classifier. SVM classified 163 of all the 778 742 lesion, the suspect is confirmed. Then cis-acting locus is characterized to pairs as TP–TP and 397 as TP–FP. The results are depicted in 779 743 decide its abnormality. This analysis using two views of same Table 4.
    780 744 breast together is referred as case-based analysis and we have Due to limiting the number of lesions per mammogram, 781 745 tried to mimic the same as explained below. two view analysis of only 860 pairs is performed which yielded 782 746 The annular region defined by the arc pair with radial with 6.81 FPs/I before classification. After geometric elimina- 783 747 distance as arc radius is used to locate the suspicious lesion on tion using arc-based method, pairing from CC to MLO yields 784 748 the MLO view which corresponds to the lesion on CC view. The 365 false pairs leading to (365/220) 1.66 FPs/I while pairing from 785 749 prediction is based on location-based properties as well as MLO to CC yields (397/220) 1.80 FPs/I. Among 220 images, the 786 750 similarity features extracted from the suspicious lesion pairs average FPs/I is reduced to 1.73. The pair is declared as likely to 787 751 on ipsilateral views. The total of 262 lesions on CC views and be malignant or benign based on classification using fusion of 788 752 276 lesions on MLO view are the candidate lesions allowed to single and two view scores as described in the next section. 789 753 undergo two views analysis. The annular or search region is Similarly, the results on DDSM dataset are depicted in Table 5. 790 754 dynamic and is based on the length of minor axis of the lesions
    755 participating in the pairing process. Hence, the missing of any 5.4. Case-based analysis 791
    756 lesion due to static parameter of region width is avoided which
    757 results in appropriate and limited pairing. There are two Once the TP–TP and TP–FP pairs are finalized, the task is to 792 758 experiments performed in this clinical study. First, all classify the pair of benign mass from the pair of malignant 793 759 suspicious lesions on CC view are selected one by one and