PSNR
Peak Signal-to-Noise Ratio, the most commonly used objective metric in image quality assessment. Measured in decibels (dB), it quantifies the pixel-level error between a reconstructed image and a reference image; higher values indicate better reconstruction quality. It is widely used for model evaluation in image generation and medical imaging.
Timeline (last 90 days)
A PSNR improvement of 1 dB is typically already significant in the NVS literature; a 4–7 dB improvement means the error is reduced several-fold.
Evaluation of general video inpainting usually directly computes whole-frame reconstruction error (PSNR, SSIM, or LPIPS), but this protocol is not suitable for video subtitle removal, because subtitles occupy only a small region of the frame and whole-frame metrics are diluted by the large amount of unmodified background pixels.
PSNR (peak signal-to-noise ratio) is calculated based on mean squared error, in units of dB, where a higher value indicates smaller pixel-level error; in the CT reconstruction literature, above 30 dB is generally considered to have good clinical usability.
When validating the quality of a density field, the PSNR or SSIM of the RGB rendering cannot be used as the sole criterion; additional density-sensitive measurement means need to be introduced, or appropriate regularization constraints imposed on the appearance capacity.
The cascading freeze incident caused end-to-end masked-region PSNR to plummet from 38.7dB to 16.7dB, a loss of 22dB.
The ClearPix team proactively removed a motion-gated safety mechanism that was already live and functioning, replacing it with a simpler anchor-comparison design, and ultimately achieved a 1.5dB improvement over the old baseline on masked-region PSNR.
The fix combining a cumulative drift budget with luminance delta compensation eliminated the recurrence of the freeze bug and stabilized quality at 38.44dB (compared to a no-frame-skip reference of 38.74dB).
The PSNR calculation formula is PSNR = 10 · log₁₀(MAX² / MSE), where MAX is typically 255 for 8-bit images.
The team set a gating rule in advance: if masked-region PSNR on any end-to-end footage falls below 36dB, the refactoring is immediately halted; this number was chosen before the team had any stake in the results.
The first run of the anchor rewrite produced split results: all four segments of the subtitle footage improved by 3–4dB (37.3/32.3/28.7/30.1dB), while the masked-region PSNR of the gradual-drift footage dropped from 38.30 to 34.69dB, falling below the stop line.
7 more timeline events
All Facts (17)
A PSNR improvement of 1 dB is typically already significant in the NVS literature; a 4–7 dB improvement means the error is reduced several-fold.
50%UnverifiedEvaluation of general video inpainting usually directly computes whole-frame reconstruction error (PSNR, SSIM, or LPIPS), but this protocol is not suitable for video subtitle removal, because subtitles occupy only a small region of the frame and whole-frame metrics are diluted by the large amount of unmodified background pixels.
50%UnverifiedPSNR (peak signal-to-noise ratio) is calculated based on mean squared error, in units of dB, where a higher value indicates smaller pixel-level error; in the CT reconstruction literature, above 30 dB is generally considered to have good clinical usability.
50%UnverifiedWhen validating the quality of a density field, the PSNR or SSIM of the RGB rendering cannot be used as the sole criterion; additional density-sensitive measurement means need to be introduced, or appropriate regularization constraints imposed on the appearance capacity.
50%UnverifiedThe cascading freeze incident caused end-to-end masked-region PSNR to plummet from 38.7dB to 16.7dB, a loss of 22dB.
50%UnverifiedThe ClearPix team proactively removed a motion-gated safety mechanism that was already live and functioning, replacing it with a simpler anchor-comparison design, and ultimately achieved a 1.5dB improvement over the old baseline on masked-region PSNR.
50%UnverifiedThe fix combining a cumulative drift budget with luminance delta compensation eliminated the recurrence of the freeze bug and stabilized quality at 38.44dB (compared to a no-frame-skip reference of 38.74dB).
50%UnverifiedThe PSNR calculation formula is PSNR = 10 · log₁₀(MAX² / MSE), where MAX is typically 255 for 8-bit images.
50%UnverifiedThe team set a gating rule in advance: if masked-region PSNR on any end-to-end footage falls below 36dB, the refactoring is immediately halted; this number was chosen before the team had any stake in the results.
50%UnverifiedThe first run of the anchor rewrite produced split results: all four segments of the subtitle footage improved by 3–4dB (37.3/32.3/28.7/30.1dB), while the masked-region PSNR of the gradual-drift footage dropped from 38.30 to 34.69dB, falling below the stop line.
50%UnverifiedViewport quality is a specific concept in VR video quality assessment, distinct from traditional video's full-frame PSNR or SSIM metrics, and requires simulating the user's actual viewing region during evaluation.
50%UnverifiedIn the field of video prediction and world models, a 0.5 dB improvement in PSNR is an extremely marginal change.
50%UnverifiedTable 4 in the paper shows that DreamDojo's PSNR metric improvement over Cosmos 2.5 is only approximately 0.5 dB.
50%UnverifiedPSNR is one of the most commonly used objective metrics in image and video quality assessment, measured in decibels, with higher values indicating that predicted frames are closer to ground-truth frames.
50%UnverifiedPSNR measures pixel-level error between a reconstructed image and a reference image in decibels on a logarithmic scale, where a 1 dB improvement typically corresponds to a perceptible improvement in detail.
50%UnverifiedAcross the four splits, the average PSNR improvement over VEG across seven scans ranged from 1.10 dB to 1.52 dB.
50%UnverifiedLPIPS (Learned Perceptual Image Patch Similarity) has a far higher correlation with human subjective evaluation than traditional PSNR and SSIM metrics.
50%