Computer-Aided Painting (Schwartz)

Math essence

Build fBm noise F : [0,1]² → [0,1] (octaves=3, persistence=0.55), normalise to zero-mean / unit-std, threshold at τ to obtain a binary mask M. For each rotation r ∈ {0, 1/K, 2/K, …, (K−1)/K × 2π}, place M_r = rot90(M, r) on a black canvas using palette index p(r) = (r + t·chromatic_shift)·K mod 6. The final image is the union (max) of the K layers.

Parameters

Reference

First 60 lines of reference.py. Full source: library_external/generative_pioneers_2/lillian_schwartz/reference.py

from __future__ import annotations

from typing import ClassVar

import imageio.v2 as imageio
import numpy as np


class LillianSchwartz:
    family: ClassVar[str] = "generative_pioneers_2"
    sketch: ClassVar[str] = "lillian_schwartz"

    def __init__(
        self,
        dim: int = 512,
        octaves: int = 3,
        persistence: float = 0.55,
        threshold: float = 0.62,
        symmetry: int = 4,
        rotate_quarters: int = 1,
        chromatic_shift: float = 0.06,
        palette_warmth: float = 0.6,
        bg_color: tuple[int, int, int] = (12, 8, 18),
        seed: int | None = None,
    ) -> None:
        self.dim = dim
        self.octaves = octaves
        self.persistence = persistence
        self.threshold = threshold
        self.symmetry = symmetry
        self.rotate_quarters = rotate_quarters % 4
        self.chromatic_shift = chromatic_shift
        self.palette_warmth = palette_warmth
        self.bg_color = np.array(bg_color, dtype=np.uint8)
        self._seed = seed
        self.rng = np.random.default_rng(seed)
        self.canvas = np.full((dim, dim, 3), self.bg_color, dtype=np.uint8)
        self.time = 0
        self._dirty = True

    def reset(self, seed: int | None = None) -> None:
        if seed is not None:
            self._seed = seed
        self.rng = np.random.default_rng(self._seed)
        self.canvas = np.full((self.dim, self.dim, 3), self.bg_color, dtype=np.uint8)
        self.time = 0
        self._dirty = True

    def _value_noise(self, w: int, h: int, period: int) -> np.ndarray:
        rng = self.rng
        gx = max(2, w // period)
        gy = max(2, h // period)
        grid = rng.random((gy + 1, gx + 1)).astype(np.float32)
        ys = np.linspace(0, gy, h, endpoint=False)
        xs = np.linspace(0, gx, w, endpoint=False)
        y0 = np.floor(ys).astype(int)
        x0 = np.floor(xs).astype(int)
        y1 = np.clip(y0 + 1, 0, gy)
        x1 = np.clip(x0 + 1, 0, gx)
        fy = (ys - y0).astype(np.float32)