A homage to Hilma af Klint, Agnes Martin, and Georgia O’Keeffe. I studied the color palettes from their works and generated palettes. It’s amazing how much you can feel the emotion just through the colors.
From Wiki page of Agnes Martin:
Agnes Bernice Martin, RCA (March 22, 1912 – December 16, 2004), was an American abstract painter. Her work has been defined as an “essay in discretion on inward-ness and silence”.
Agnes Martin, Happy Holiday 1999
I used k-means clustering to find the color palettes of artists I liked, by specifying the number of k from 2 to 10. Then I randomly select a palette and draw color strips based on the frequency of the color in the original artwork. Some white noise is added in the end and pixels randomly swapped for a more organic look.
def generate_palettes_from_img(img, theme=None, nmin=2, nmax=10): results =  cols = ['theme', 'n', 'r', 'g', 'b', 'freq_start', 'freq_end'] for n in range(nmin, nmax): hist, cluster_centers, clt = kmeans_colors(img, n) accu_freq = 0 for i in range(len(cluster_centers)): row = [theme, n] freq = hist[i] cluster_center = np.round(cluster_centers[i], 2) row.extend(list(cluster_center)) row.extend([accu_freq, accu_freq + freq]) accu_freq += freq results.append(row) plot_palette(hist, cluster_centers) df = pd.DataFrame(results, columns = cols) return df
Here is an example for Paul Klee’s work’s palette.
k-means clustering for palette