Nine types of image narrative textures

model

gpt-image-2

Generate a set of nine equally divided image narratives for the cats in the reference image: the same theme core repeatedly appears in each unit, but there are slight changes in posture, gaze, expression, local movements, or symbolic states, forming a continuous rhythm like an emotional sample book. The main body bears the structural weight with a clean grayscale or low chroma image texture, while the background remains bright, flat, and ventilated, with clear boundaries without excessive decoration; Overlay a layer of free lines, such as crayons, oil pastels, or children's hand drawn traces, on each unit to create a direct collision between the hand drawn layer and the image layer. The hand drawn content must be derived from the emotions, materials, symbols, and narrative clues of the theme itself, surrounding the subject's head, contours, blank areas, and edges with circular drawing, framing, smudging, simple pen symbols, short line segments, or scattered handwritten shapes, which are both like instant graffiti and can enhance the personality differences of each grid. Colors are extracted from the material, season, cultural temperament, and emotional temperature of the theme itself. A large area of bright neutral background and clear grayscale are responsible for the order of the main body. A small amount of highly saturated theme colors are distributed as emotional emphasis in the hand drawn layer, maintaining high brightness, cleanliness, lightness, and liveliness. The color scale is clear, avoiding dirty, outdated, and cloudy coverage. The overall preservation of rough hand drawn particles, indentations, broken lines, and imperfect edges, the image reads first with a strong order of repetitive matrices, then broken by different naive graffiti in each grid, ultimately presenting a bright, playful, direct, and handmade visual system.

parameters

Reference Images