Computer Methods And Programs In Biomedicine[JOURNAL]

Sun 200 papers

PPG-basis: A python-based toolbox for fast photoplethysmogram decomposition and reconstruction.

Putcha A, Keyal R, Piasecki L … +5 more , Sharma A, Shmuylovich L, Lou Y, Kosorok M, Bai W

Comput Methods Programs Biomed · 2026 Aug · PMID 42105444 · Full text

BACKGROUND & OBJECTIVE: Photoplethysmogram (PPG) data are commonly used for critical biometric extractions, such as heart rate and blood oxygenation. However, PPG data contains valuable information on other metrics, such... BACKGROUND & OBJECTIVE: Photoplethysmogram (PPG) data are commonly used for critical biometric extractions, such as heart rate and blood oxygenation. However, PPG data contains valuable information on other metrics, such as vascular stiffness, within the morphology of the data itself, rather than individual points. As such, a growing body of literature is focused on reliably decomposing PPG data morphology for subsequent biometric extraction. Separately, the accurate reconstruction of noisy PPG data for downstream model training is necessary for developing noise-resistant models. However, to our knowledge, there is no unified toolbox that enables the testing of multiple decomposition methods and subsequent reconstruction. METHODS: We reformulate the PPG generation problem in the phase, rather than time, domain and use a look-up table to avoid recompilation. We then provide three distinct kernels to represent PPG data and develop a two-stage decomposition pipeline to accurately reconstruct PPG data. RESULTS: We demonstrate an order-of-magnitude decrease in computational time using our phase-domain approach while retaining high accuracy when compared to existing methods. We further demonstrate the utility of this work through the decomposition and reconstruction of experimental data with injected noise, as well as the extraction of complex biometrics in N = 50 subjects from the MIMIC-III clinical database. CONCLUSION: On clean input, ppg-basis recovers morphological features - such as dicrotic notch depth and augmentation index - with substantially lower error than bandpass filtering alone, despite higher global waveform MSE. Under noise injection, the framework maintains stable morphological feature detection across SNR levels, whereas filter-based approaches produce artefactual detections that inflate with increasing noise. Furthermore, we anticipate combining morphologically accurate PPGs with noise-injection frameworks could be used for developing noise-resistant models and testing the robustness of existing models focused on PPG analysis.

Free PDF PubMed ↗

IML-UNet: A brain-inspired spatiotemporal collaborative encoding method for medical image sequence registration.

Liu X, Chen X, Wang Z … +1 more , Wei Y

Comput Methods Programs Biomed · 2026 Aug · PMID 42102639 · Publisher ↗

BACKGROUND AND OBJECTIVE: Deformable medical image registration is important for radiotherapy planning, respiratory motion analysis, and organ function assessment. In medical image sequences such as 4D CT, organ motion i... BACKGROUND AND OBJECTIVE: Deformable medical image registration is important for radiotherapy planning, respiratory motion analysis, and organ function assessment. In medical image sequences such as 4D CT, organ motion involves large displacements, local non-rigid deformations, and continuous temporal changes. Registration requires accurate spatial matching together with temporal consistency and deformation stability across the sequence. However, most existing methods either process spatial and temporal information in separate stages or treat 4D registration as independent pairwise tasks, limiting spatiotemporal interaction and long-range temporal dependency modeling. We therefore propose a spatiotemporal joint registration method that models multi-scale spatial information and temporal dependencies within a unified framework to improve accuracy, stability, and sequence consistency under large-displacement conditions. METHODS: We propose IML-UNet, a deep learning-based method centered on a new recurrent cell, Inception MLP-like LSTM (iMLSTM), inspired by spatiotemporal collaborative encoding in partially overlapping neuronal populations. The iMLSTM integrates parallel multi-scale feature extraction with long-range temporal modeling to form fused spatiotemporal representations at each network level and reduce hierarchical feature degradation. A Progressive Frequency Architecture (PFA) flexibly adjusts the contributions of high- and low-frequency features across network levels to strengthen representation capability. During training, an Adaptive Weighted Multi-Scale Structural Similarity Index Measure (AM-SSIM) dynamically optimizes multi-scale similarity assessment. We also design efficient sequence registration strategies to capture dynamic features of 4D sequences and support diverse registration scenarios. RESULTS: On two publicly available 4D CT datasets, our method achieved TREs of 1.09 mm and 0.81 mm, respectively, demonstrating superior registration performance over prior methods. CONCLUSIONS: The proposed method delivers stable and accurate registration for lung 4D CT image sequences, with its main advantage lying in hierarchical intra-level spatiotemporal coupling rather than in the simple stacking of spatial and temporal modules. The observed performance gains are achieved with comparable model size and inference time, which strengthens the practical relevance of the method. It provides a promising framework for modeling complex spatiotemporal deformations in this setting. However, the current evidence is limited to a per-case optimization setting, and broader cohort-level generalizability remains to be further validated.

Computer Methods And Programs In Biomedicine[JOURNAL]

About