Table 1 Summary of Cervical Cancer Detection Literature.
From: A deep ensemble learning approach for squamous cell classification in cervical cancer
Reference | Method | Dataset | Results | Remarks |
|---|---|---|---|---|
Mango et al.11 | Pap smear test + ANN model | N/A | N/A | Detection of cancerous cells in cervix. |
Sukumar and Gnanamurthy12 | MRI scans + SVM + NN | Herlev data | 99.1% acc. in 2-class | Automated diagnosis using MRI scans. |
Bora et al.13 | CNN-based classification | Private dataset | Improved accuracy with feature selection | Deep CNN for image identification. |
Hyeon et al.14 | CNNs + VGG16 for feature extraction | 7134 MRIs | SVM’s F1 score superior | Classifying cervical MRIs as normal or infectious. |
Promworn et al.30 | Comparative analysis of models | N/A | DenseNet161 achieved 94.38% acc. | DenseNet161 excelled among five models. |
ColpoNet15 | Inspired by DenseNet | Nat. Cancer Institute dataset | Accuracy of 81.353% | Based on computationally efficient DenseNet. |
Parikshit Sanyal et al.16 | CNN for detecting ’abnormal’ foci | 1838 microphotographs | 95.46% diagnosis acc. | High accuracy in classifying normal and abnormal foci. |
Karunakaran et al.17 | Ultrasensitive SERS for sample prediction | Cervix cell samples | Average acc. of 95.46% | Predicting normal, HSIL, and CSCC. |
Taha et al.18 | Pre-trained CNN architecture | Herlev dataset | 99.19% acc. in 2-class | Effectiveness of pre-trained CNN architecture. |
Kudva et al.19 | Hybrid transfer learning system | AlexNet and VGG-16 features | Classification acc. of 91.46% | Improved classification with focused filters. |
Xue et al.20 | Ensemble Transfer Learning (ETL) | Herlev dataset | Highest acc. of 98.61% | ETL after developing multiple deep learning models. |
Chen et al.21 | Fine-tuned CNN architectures | 4993 histology images | Achieved 97.42% classification acc. | Effectiveness of transfer learning for histopathology images. |
Ghoneim et al.22 | CNN-based approaches with ELM classifiers | Herlev database | 99.5% detection acc. and 91.2% classification acc. | Utilized ELM classifiers for deep-learned characteristics in cell images. |
Kang et al.25 | Raman spectroscopy, H-CNN | Tissue samples | Over 94% accuracy in classifying tissues | H-CNN promising for precise cervical cancer diagnosis |
Youneszade et al.26 | - | Review of techniques, architectures, and segmentation methods | Overview of DL in cervical cancer screening | Emphasizes the need for further research |
Pacal et al.27 | ViT, CNN-based models, | Massive dataset | Record-breaking classification accuracy | Potential for early diagnosis and reduced mortality rates |
Pramanik et al.28 | Fuzzy distance-based ensemble | Pap smear images | Promising initial results in accuracy and efficiency improvement | Further research needed for full assessment |
