Table 1 Summary of important existing models for CC subtype classification.

BiNext-Cervix⁵

ConvNext and BiFormer models

SipakMed

83.51%

Progressive Resizing approach + PCA⁵⁷

Extracts features using ResNet-152 and VGG-16 with progressive resizing (224 × 224 → 1024 × 1024); PCA for dimensionality reduction + Majority voting based classification (SVM + RF)

SipakMed + LBC

98.47%

CTCNet⁵⁸

CNNs and Transformers. Deformable Large Kernel Attention (DLKAttention)

SipakMed

97.74%

6 Deep learning models + SVM⁶¹

VGG16, Xception, DenseNet169, InceptionV3, ResNet101, and Inception ResNet + SVM for classification

SipakMed

95.66%

Improved CervicalNet⁶²

U-Net for segmentation + GCN for classification

SipakMed + Herlev

Accuracy-98.61% precision-97.33%, specificity-97.12%, recall- 97.11%, F1-score-97.56%

MaxCerVixT⁶³

CNN-based ViT

SipakMed

99.02%

TL based CNN⁴⁰

TL based EfficientNet B3 and progressive resizing

SIPaKMed

99.70%

CervixFormer³⁶

Swin Transformer

SipakMed, and Cervix93

SipakMed-98.29% Cervix93-97.01%

VisionCervix⁴⁵

Vision Transformer (ViT) and fine-tuned MobileNet

SipakMed

Accuracy- 97.65%, precision-99.54%, recall- 97.65%, f1 score- 98.58%

CVM-Cervix²⁰

Xception model, ViT, MLP

CRIC, SipakMed

Accuracy- 92.87%, precision-92.80%, recall- 92.90%, f1 score- 92.80%

CNN³⁹

CNN

SipakMed

Accuracy-91.13%

CACCD-GOADL⁴⁸

MobileNetv3 model with Gazelle Optimizer Algorithm (GOA)

Herlev

f1 score- 95.71%, Accuracy- 98.69%, recall- 95.37%, precision-95.24%

CytoBrain⁵⁵

CompactVGG

Cervical cancer WSI images

Accuracy-88.30%, specificity-91.03%, f1 score-87.04%, Sensitivity-92.83

SOD-GAN + F-SAE⁵²

Fine-tuned Stacked Autoencoder based GAN

Colonoscopy images

Accuracy-94.8%

GCN⁵⁴

Graph Convolution Network

SipakMed

Accuracy-98.37%

CerCanNet⁴⁷

ResNet18 + Quadratic Support Vector Machine

SipakMed

Accuracy-96.3%

CerviXpert⁵¹

Customised CNN

SipakMed

Accuracy for three class classification-98.04%, Five class classification- 98.60%