Files
BGfilter/configs/default.yaml
T
lhk229 639fdf01b6 Gate chroma suppression by the matte's own confidence
suppress_alpha_by_chroma was the one stage where colour evidence could veto the
matting model outright, and every observed failure had the same shape: ViTMatte
was right (raw ~0.97 on a white shirt over pastel blue, on blush over pink) and
the bgc-keyed suppression crushed it, because bgc magnitude cannot tell a
background-coloured subject from background residue. The matte's raw alpha can:
it rates enclosed background pockets low (~0.28 median) and subject texture high.

Suppression is now scaled by (1 - smoothstep(raw_alpha, suppress_raw_lo,
suppress_raw_hi)) (defaults 0.85/0.98): colour evidence only vetoes the matte
where the matte itself is unsure. The gate applies only when alpha really comes
from ViTMatte; chroma-seeded alpha passes raw_alpha=None and keeps old behaviour.
Set suppress_raw_lo: 1.0 to disable (verified byte-identical to the old output).

Validated on both screen colours: pink #F4D8E8 face restored (skin alpha 0.841 ->
0.993, holes 17.2% -> 0.6%), pastel blue #CFEFFF unchanged-to-better (shirt +0.01,
wisps identical, gap clearing 99.8%). Also fixes the long-deferred chroma-only
pipeline defects (white shirt 0.900, light skin 0.989). CLI output verified
byte-identical to the reviewed prototype on both images.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-03 23:54:28 +08:00

73 lines
2.5 KiB
YAML

# Background-colour prior. null = auto-detect the flat background colour from the
# image border (raises if the border is not one clean flat colour). A hex string
# like "#CFEFFF" sets it explicitly. Also overridable via --screen-color.
screen_color: null
model:
model_name: hustvl/vitmatte-base-composition-1k
device: cpu
matting_method: vitmatte
fallback_to_chroma_alpha: false
segmentation:
# enabled: true = single-segmenter pipeline (one seg model drives the trimap)
# enabled: false = no seg model, chroma-only pipeline (Chroma + ViTMatte)
enabled: true
backend: birefnet # single trimap segmenter; switchable at runtime: --seg-backend birefnet|anime-seg
model_name: ZhengPeng7/BiRefNet # matches backend; --seg-backend anime-seg auto-selects skytnt/anime-seg
device: cpu
input_size: 1024
trimap:
sure_bg_threshold: 0.92
sure_fg_threshold: 0.12
unknown_radius_ratio: 0.012
fg_safe_radius_ratio: 0.006
seg_core_threshold: 0.60
seg_loose_threshold: 0.08
# Trimap mode (segmentation pipeline only; also settable via --trimap-mode):
# directional = chroma + seg + hue split; background-hued residual
# (blue between hair) -> unknown, white shirt kept. (default)
# seg = original fuse_trimap (no hue split).
# directional-hard-bg = hard-remove background-hued pixels (aggressive; can eat
# cool/shadowed white cloth).
mode: directional
seg_low: 0.15
bg_hue_proj_min: 4.0
alpha_post:
chroma_suppress: true
chroma_suppress_bg_low: 0.35
chroma_suppress_bg_high: 0.80
chroma_suppress_strength: 1.0
# Matte-confidence gate: suppression fades where ViTMatte's raw alpha is high
# (ramp lo -> hi), so colour evidence only vetoes the matte where it is unsure.
# Set suppress_raw_lo: 1.0 to disable.
suppress_raw_lo: 0.85
suppress_raw_hi: 0.98
foreground:
enabled: true
method: ml
fallback_to_unmix: true
ml_regularization: 0.00001
# Legacy "unmix" method params (used only when method: unmix).
edge_low: 0.005
edge_high: 0.995
min_unmix_alpha: 0.08
unmix_strength: 0.70
local_strength: 0.45
local_blur_radius: 11
local_alpha_threshold: 0.92
local_bg_confidence_max: 0.20
green_excess_margin: 0.015
bg_confidence_weight: 0.70
despill:
enabled: true
edge_low: 0.005
strength: 0.92
edge_expand_radius: 2
alpha_weight_floor: 0.35
color_margin_lab: 4.0 # Lab chroma margin along the background-colour direction