tbcm 0.4

Ten Bit Color Maps

Ten Bit Color Maps

Features

• 1024 colors in each sequential colormap
• Perceptually uniform - each color is same perceptual distance from the previous color (before rounding)
• Raw color data rounded so they can be used in a byte tensor (see pytorch example below)
• All maps start at black

Installation

pip install tbcm

Usage

With matplotlib

from tbcm import tb_inferno, tb_oleron

then to use tb_inferno, do:

import matplotlib.pyplot as plt
import numpy as np
x,y = np.meshgrid(np.linspace(-1,1,15),np.linspace(-1,1,15))
z = np.cos(x*np.pi)*np.sin(y*np.pi)

fig = plt.figure(figsize=(9,4))
ax1 = fig.add_subplot(121, projection='3d')
ax1.plot_surface(x,y,z,rstride=1,cstride=1,cmap=tb_inferno)
ax2 = fig.add_subplot(122)
cf = ax2.contourf(x,y,z,51,vmin=-1,vmax=1,cmap=tb_inferno)
cbar = fig.colorbar(cf)

result:

and the same with tb_oleron is:

fig = plt.figure(figsize=(9,4))
ax1 = fig.add_subplot(121, projection='3d')
ax1.plot_surface(x,y,z,rstride=1,cstride=1,cmap=tb_oleron)
ax2 = fig.add_subplot(122)
cf = ax2.contourf(x,y,z,51,vmin=-1,vmax=1,cmap=tb_oleron)
cbar = fig.colorbar(cf)

result:

With pytorch

First import the color data as lists and convert to byte tensors:

from tbcm import tb_inferno_data, tb_oleron_data

import torch
import numpy as np
from PIL import Image

tb_oleron = torch.tensor(tb_oleron_data, dtype=torch.uint8)
tb_inferno = torch.tensor(tb_inferno_data, dtype=torch.uint8)

Then define this mapping function to convert data to a range appropriate for this colormap:

def map_values(data, cm_range=(0.126, 0.99), invalid=None):
    if invalid is None:
        hist, bin_edges = torch.histogram(data.view(-1), bins=int(1e4))
    else:
        hist, bin_edges = torch.histogram(data[~invalid], bins=int(1e4))
    bin_centers = (bin_edges[:-1] + bin_edges[1:]) / 2
    cdf = hist.cumsum(dim=0)
    cdf = cdf / cdf[-1]
    out = torch.from_numpy(np.interp(data.numpy(), bin_centers.numpy(), cdf.numpy()))
    scale = cm_range[1] - cm_range[0]
    if invalid is None:
        out = out - out.min()
        out = out * scale / out.max()
    else:
        out = out - out[~invalid].min()
        out = out * scale / out[~invalid].max()
    out = out + cm_range[0]
    return out

Finally use the tb_oleron colormap like:

z = torch.cos(torch.linspace(-1, 1, 256).view(1, -1) * torch.pi) * torch.sin(
    torch.linspace(-1, 1, 256).view(-1, 1) * torch.pi
)

h, w = z.shape
idx = map_values(z).view(-1).mul(1023).round().clamp(0, 1023).long()
img = tb_oleron[idx].view(h, w, 3)
Image.fromarray(img.numpy())

Result: