V4 - Frame-wise Flicker Removal

Date: 2026-07-13. Source video: /local/scratch-2/fy277/temp/vdp_screening/input/cycle_view.mp4. Clip: 3.0s at 8 fps, 384x256. Optimization iterations: 200 for each correction run. Correction model: per-frame per-channel affine tensor.

0 - Label Legend

CVVDP usage: the whole video clip is passed to CVVDP to compute JOD and, for cvvdp_hm losses, a detached raw CVVDP heatmap bank. Optimization does not backpropagate through the CVVDP scalar here; it optimizes L1 or heatmap-weighted L1.

alpha_only: learn per-frame, per-channel multiplicative gain only: out_t = alpha_t * input_t.
alpha_beta: learn per-frame, per-channel gain and bias: out_t = alpha_t * input_t + beta_t.
cvvdp_hm_a1: detached CVVDP heatmap weighted L1 with heatmap weight strength 1.
cvvdp_hm_a4: detached CVVDP heatmap weighted L1 with heatmap weight strength 4.

1 - Flicker Inputs

InputJODPSNRSSIML1LPIPS
cycle_view_global_brightness_amp0p206.33442.1140.9820.045680.0072
cycle_view_color_flicker_amp0p502.34021.5240.9180.083990.1618

2 - Results

cycle_view_global_brightness_amp0p20

ModelLossJODPSNRSSIML1LPIPS
alpha_onlycvvdp_hm_a19.77464.0260.9990.000880.0002
alpha_onlyl19.66763.6300.9990.000970.0003
alpha_onlycvvdp_hm_a49.70163.2320.9990.001140.0003
alpha_betacvvdp_hm_a49.16858.2130.9990.004020.0013
alpha_betal19.31558.1290.9990.004070.0011
cycle_view_global_brightness_amp0p20 input
alpha_only - cvvdp_hm_a1
alpha_only - l1
alpha_only - cvvdp_hm_a4
alpha_beta - cvvdp_hm_a4
alpha_beta - l1

cycle_view_color_flicker_amp0p50

ModelLossJODPSNRSSIML1LPIPS
alpha_onlycvvdp_hm_a18.89241.0510.9910.003400.0112
alpha_onlycvvdp_hm_a48.88140.9880.9910.003360.0113
alpha_onlyl18.87740.8600.9910.003340.0113
alpha_betacvvdp_hm_a48.31139.2800.9910.006430.0118
alpha_betal18.70838.8460.9910.006120.0120
cycle_view_color_flicker_amp0p50 input
alpha_only - cvvdp_hm_a1
alpha_only - cvvdp_hm_a4
alpha_only - l1
alpha_beta - cvvdp_hm_a4
alpha_beta - l1

3 - Verdict

Cycle-view flicker-removal rerun. The synthetic inputs are generated from the new reference clip, then a per-frame affine correction is fit with L1 or detached CVVDP-heatmap-weighted L1. All embedded videos are H.264/yuv420p/faststart.

4 - Config

{
  "source_video": "/local/scratch-2/fy277/temp/vdp_screening/input/cycle_view.mp4",
  "work_dir": "/local/scratch-2/fy277/temp/vdp_screening/work/v1_v4_cycle_view_20260713",
  "v1_report": "/local/scratch-2/fy277/temp/vdp_screening/reports/v1_compression_2026-07-12_cycle_view_new",
  "v4_report": "/local/scratch-2/fy277/temp/vdp_screening/reports/v4_flicker_removal_2026-07-12_cycle_view_new",
  "seconds": 3.0,
  "start": 0.0,
  "fps": 8,
  "width": 384,
  "v4_iters": 200,
  "only": "all",
  "height": 256,
  "frames": 24
}