Table 5.
Comparison of PFP-LHCINCA model with published results of other ultrasound-based fetal sex classification methods.
| Study | Method | Dataset | Best accuracy (%) |
|---|---|---|---|
| Maysanjaya et al. [7] | Learning vector quantization, artificial vector quantization | 64 males 25 females |
63.0% |
|
| |||
| Aljuboori et al. [8] | Fuzzy C-mean, discrete wavelet transform, local binary pattern, median, Laplacian filters | 50 males 50 females |
94.0% |
|
| |||
| PFP-LHCINCA | Pyramidal fixed-size patch division, local phase quantization and histogram of oriented gradients based feature extraction, hybrid Chi2 and iterative neighborhood component analysis feature selection | 339 males 332 females |
99.11% (kNN classifier tuned with Bayesian optimizer) |