Table 5 Comparison of the proposed model with state-of-the-art literature.
Author | Year | Dataset type | Technique | Use of XAI | Accuracy | Misclassification rate |
|---|---|---|---|---|---|---|
Chiang et al.47 | 2003 | 105 healthy controls and 151 patients with calcium oxalate stones | ANN and DA | No | ANN: 89% and DA: 75% | ANN: 11% and DA: 25% |
Dussol et al.48 | 2006 | 119 stone formers and 96 controls | ANN (LDA and MVDA) | No | LDA: 75.8% and MVDA: 74.4% | LDA: 24.2% and MVDA: 25.6% |
Cauderella et al.49 | 2011 | 80 patient’s data | ANN and LR | No | ANN: 88.8% and LR: 67.5% | ANN: 11.2% and LR: 32.5% |
Kumar and Abhishek50 | 2012 | Data from 1000 patients | LVQ, MLP, and RBF | No | LVQ: 84%, MLP: 92%, and RBF: 87% | LVQ: 16%, MLP: 8% and RBF: 13% |
Ebrahimi and Mariano51 | 2015 | KUB CT scan slides from 39 patients | Image processing techniques and geometry principles | No | 84.61% | 15.39% |
Kazemi and Mirroshandel52 | 2018 | Numeric characteristics from 936 patients | Ensemble learning model | No | 97.1% | 2.9% |
Li and Elliot53 | 2019 | 1874 CT KUB reports | NLP | No | 85% | 15% |
De Perrot et al.54 | 2019 | 416 patient data | ML model | No | 85.1% | 14.9% |
Kahani et al.55 | 2020 | KUB x-ray images | LASSO with ML classifiers | No | 96% | 4% |
Jungmann et al.56 | 2020 | 1714 LDCT images | NLP | No | 72% | 28% |
Annameti Rohith et al.57 | 2021 | 114 ultrasound images | Median and rank filters | No | Median filter: 86.4%, and rank filter: 82.2% | Median filter: 13.6%, and rank filter: 17.8% |
Suresh and Abhishek58 | 2021 | KUB ultrasound images | Image processing techniques | No | 92.57% | 7.43% |
Jendenber et al.59 | 2021 | NCCT images of 341 patients containing a distal ureteral stone, phlebolith, or both | CNN | No | 92% | 8% |
Cui et al.60 | 2021 | 625 CT images | DL and threshold-based model | No | 90.30% | 9.70% |
Yildirim et al.61 | 2021 | 1799 coronal CT images | XResNet-50 | No | 96.82% | 3.18% |
Tsitsiflis et al.62 | 2022 | Medical data of 716 patients | ANN | No | 81.43% | 18.57% |
Valencia et al.63 | 2022 | CT scans of around 40 patients | Image processing | No | 92.5% | 7.5% |
Proposed model | 2024 | 14,265 KUB x-ray images | VGG16 empowered with LRP | Yes | 97.41% | 2.59% |
