Orbital floor fractures may be overlooked on computed tomography (CT) because fracture lines can exhibit low contrast and indistinct edges. This study presents a hybrid discrete wavelet transform–modified bat algorithm (DWT–MBA) framework as a technical feasibility approach for enhancing the conspicuity of pure orbital blow-out fracture margins while preserving anatomical structure. Each representative CT slice is decomposed into multiscale DWT sub-bands, and only the high-frequency coefficients (LH, HL, HH) are adaptively modified. An application-specific modified bat algorithm is used to optimize a hard-threshold scaling factor \(\alpha \in [0.1,1.0]\) through an objective that balances fidelity, total-variation regularization, and structural similarity (SSIM), with entropy-guided parameter adaptation to improve robustness in low-contrast CT. Five wavelet families (Haar, Daubechies-2, Daubechies-4, Coiflet-2, Coiflet-4) were evaluated. In a feasibility cohort of eight representative coronal CT slices (one per patient), DWT–MBA achieved PSNR up to 51.3 dB, SSIM up to 0.99, FSIM up to 0.99, and entropy up to 5.78, with generally higher fidelity- and entropy-related metrics than five baseline enhancement methods. Because these metrics were computed against the original slice as reference, they should be interpreted primarily as indicators of fidelity and anatomical consistency rather than clinical diagnostic benefit. Qualitative inspection suggested potentially increased fracture-boundary conspicuity on representative 2D images; however, clinical utility remains to be established through blinded radiologist reader studies and larger multi-slice or volumetric validation. Overall, the findings support the technical feasibility of entropy-guided wavelet-domain optimization for low-contrast orbital CT slice enhancement.
