Calibrated per-pixel uncertainty for low-dose paediatric chest-radiograph denoising at no fidelity cost
Year: 2026
Type: preprint
Abstract: Background. Deep denoisers restore low-dose chest radiographs but emit a single point estimate with no indication of where the output is reliable, so a confident-looking reconstruction can be locally wrong. In paediatric radiography, where the radiation-dose imperative is sharpest, we ask whether a per-pixel uncertainty map can be attached at no meaningful cost to reconstruction quality. Methods. We develop PP-VAE-Hformer, a hybrid convolution–transformer denoiser with a variational-autoencoder bottleneck (epistemic uncertainty map) and a heteroscedastic dual head (aleatoric map), and run a 19-arm ablation isolating five composite-loss terms on Poisson–Gaussian-degraded paediatric chest radiographs (Kermany collection; 624-image held-out test set), against eight discriminative baselines retrained on identical data and noise. Calibration is assessed by reliability diagrams and post-hoc σ-scaling; comparisons use Bonferroni-corrected Welch tests with Cohen’s d as the primary discriminant. Results. A heteroscedastic negative-log-likelihood (NLL) objective buys a calibrated aleatoric map for 1.2 dB PSNR, which two-stage fine-tuning recovers to a statistically indistinguishable 0.010 dB. For the variational variant the Kullback–Leibler (KL) annealing schedule, not the bottleneck itself, is decisive: cyclic annealing closes a 1.10 dB gap over posterior-collapsing linear warmup. Predicted uncertainty tracks realised error (Pearson r > 0.8) with an optimal recalibration scale within 1% of unity; the aleatoric and epistemic maps are only moderately correlated (¯ r = 0.53), flagging complementary regions. We additionally document a previously unreported destructive interaction between structural-similarity and Sobel-edge supervision at literature-default weights. Conclusions. A calibrated per-pixel uncertainty map can be attached to a paediatric low-dose CXR denoiser at effectively no reconstruction cost. The near-perfect aleatoric calibration partly reflects the matched (“inverse-crime”) noise model, positioning the epistemic map as the genuine reconstruction-uncertainty signal; a blinded reader study and real low-dose data are the immediate next steps.
Source: Research Square
Authors: Samuel Tekle Mengistu, Kyriakos Flouris
Institutions: University of Oxford, University of Cambridge
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7213349920