|
| S. no. | Authors | Algorithms used | Merits | Demerits/future work |
|
| (1) | Dorgham et al. | MBO | The framework is more reliable and effective, applying multistage threshold values for image fragmentation | An effective balance for searching at local and global points needs to be addressed |
| (2) | Abualigah et al. | DATA | Improved local search among the tested standard methodologies | In the future, this methodology can be applied to many industrial and text processing techniques |
| (3) | Sun et al. | DeepMedic+CRF, Ensemble+CRF | Provided a complete study about the previously used algorithms and did a contrast survey that could guide beginners to research this topic | In the future, an efficient hybrid method could be used for optimal differentiation |
| (4) | Tamoor et al. | Gaussian-based spatially varied contour | A broad penalty term and distance validation terms were incorporated for smooth evolution | It could be experimented with in applying to other types of issues as well |
| (5) | Sapna Juneja | CNN, ReLU | Least dimensionality issue | DL techniques could apply by choosing more data attributes and computing a whole lot of information |
| (6) | Ramos-Soto et al. | MCET-HHO multilayer method | The developed framework is claimed to outperform the standard procedure and shows an incredible analysis of visuals in detail | In the future, reinforcement algorithms could be applied |
| (7) | Bataineh et al. | KNN, ANN, and SVM | Increased accuracy | The ANN approach might have been driven under overriding. In the future, the developers plan to extend the research by intaking more features and working on larger-scale data |
| (8) | Haleem et al. | Bi-LSTM | The states that the developed method reduced work for domain experts with calculating necessary data with a simultaneous approach | 100% accuracy may show a machine overriding issue and ignorance in dimensionality |
|
