Table 2.
Short list of some non-deep approaches for breast lesion segmentation. For each entry, the table reports the publication year, whether it uses a MCT, the used set of features, the approach category, and the obtained results, as reported by the authors (ACC: voxel-based accuracy; AUC: area under the ROC curve; DR: detection rate; DSC: Dice similarity coefficient; HD: voxel-based Hausdorff distance; OR: overlap ratio; SEN: voxel-based sensitivity).
| Study | Year | MCT | Features | Approach | Performance |
|---|---|---|---|---|---|
| Agner et al. [14] | 2009 | DYN | MODEL, MORPH | HD 11.57 | |
| Bhooshan et al. [15] | 2010 | DYN | MODEL | AUC 0.83 | |
| Cai et al. [16] | 2014 | DYN | MODEL, MORPH | AUC 0.93 | |
| Dalmis et al. [17] | 2016 | DYN | MORPH | AUC 0.85 | |
| Fusco et al. [18] | 2012 | DYN, GEO | MODEL | ACC 0.91 | |
| Hassanien et al. [19] | 2012 | TEX | MODEL | ACC 0.98 | |
| Jayender et al. [20] | 2014 | DYN | MODEL | DSC 0.77 | |
| Lee et al. [21] | 2010 | ✓ | DYN | MODEL | AUC 0.88 |
| Marrone et al. [22] | 2013 | DYN | MODEL | ACC 0.98 | |
| McClymont et al. [23] | 2014 | ✓ | DYN | MODEL, MORPH | DSC 0.76 |
| Moftah et al. [24] | 2014 | DYN | MODEL | ACC 0.89 | |
| Nagarajan et al. [25] | 2013 | DYN | MODEL | AUC 0.82 | |
| Vignati et al. [26] | 2009 | ✓ | DYN | FILT | SEN 0.93 |
| Vignati et al. [27] | 2011 | ✓ | DYN | FILT | DR 0.89 |
| Wang et al. [28] | 2013 | ✓ | DYN | MODEL | OR 0.93 |
| Wang et al. [29] | 2014 | DYN | MORPH | ACC 0.91 | |
| Zheng et al. [30] | 2009 | DYN | MORPH | ACC 0.97 |