Loss Functions Analysis of Performance Improvements in Single-Image Super-Resolution
Bibliographic record
Abstract
The building rooftop delineation plays a critical role in urban planning and management. Current methods on improving the delineation accuracy of the building rooftops mostly focused on the data improvements and methodology improvement. However, as the current deep learning networks have already reached high accuracy, the data quality has then become the important topic. Although loss functions serve the purpose of quantifying reconstruction errors and directing the optimization process of the model in super-resolution (SR), limited research related to their impact on SR were implemented. In this study, we focused on improving the spatial resolution of the building datasets by investigating numerous loss functions, including the Mean Absolute Error (MAE) loss function, the Mean-Squared Error (MSE) loss function, the SmoothL1 loss function and the Charbonnier loss function. With our proposed Single-Image Super-Resolution (SISR) network, Residual Feature Aggregation- Pyramid Vision Transformer- Involution Network (RFA-PVTInvoNet) and other typical SISR networks, the loss functions were applied to further enhance the performance. By using Peak Signal-Noise Ratio (PSNR) and Similarity Structure Index Measurement (SSIM) as the evaluation metrics, the RFA-PVTInvoNet with Charbonnier loss function showed the highest performance compared with other approaches, with PSNR of 22.046 dB and SSIM of 0.502 on the WHU Building Dataset, which demonstrate the superior performance of the Charbonnier loss in SISR.
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.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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 itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".