Upconversion Photoluminescence Lifetime Imaging via Multi-Prior Physics-Enhanced Deep Learning
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Upconversion photoluminescence lifetime imaging has emerged as a transformative analytical tool by leveraging the temporal signatures of luminescent probes to resolve molecular interactions. While compressed ultrafast photography offers distinct advantages for high-speed imaging, its utility in quantitative applications, such as food safety, remains constrained by reconstruction artifacts, particularly spatial distortion and edge detail loss. To address these limitations, we present a method of single-shot compressed upconversion photoluminescence lifetime imaging (sCUPLI) empowered by a multi-prior physics-enhanced neural (mPEN) network. This framework synergizes physical models of photoluminescence dynamics and imaging processing, subpixel information loss of the compressed encoding, extended sampling priors, sparsity constraints, and deep learning to achieve precise quantification of upconversion photoluminescence lifetimes. To demonstrate its potential in beverage additive detection, we apply mPEN–sCUPLI to analyze synthetic colorants using rare-earth-doped upconversion nanoprobes. The system excels at resolving subtle optical heterogeneities in beverage matrices, achieving superior spatiotemporal resolution and artifact suppression compared to conventional methods. It attains a spatial resolution of 90.5 lp/mm at a frame rate of 33,000 fps, achieving a spatial resolution improvement of approximately 3.56 times. Quantitative analysis shows that mPEN–sCUPLI reconstruction improves the average peak signal-to-noise ratio by 4 dB and enhances the sharpness and fidelity of imaging by 1.85 times. Given its ability to achieve high-fidelity and high-throughput photoluminescence lifetime imaging from single-shot 2-dimensional measurements, the proposed mPEN–sCUPLI approach is expected to be widely adopted in food safety detection applications.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it