Pôster de Viagem com Memória de Cidade em Giz de Cera Infantil

Create a complete, standalone 3:4 vertical before-and-after travel illustration poster. The top half preserves the input Italian travel street photo with high fidelity, keeping the people, clothing, poses, architecture, street, lighting, colors, perspective, and subject placement unchanged, without any illustration treatment. The bottom half redraws the same people and the same street as a "playful crayon city memory." Only extract city details that are truly visible and recognizable in the original photo, such as narrow alley perspective, warm-colored walls, windows, street lamps, stone pavement, festive lights, railings, or a map in the tourist's hands; choose based on each photo's own content, without applying a fixed mountain view, car, or generic landmark. Preserve the number of main people, their relative positions, poses, direction of movement, clothing silhouettes, and the relationship between the people and the street. The people may be simplified in a childlike way, but they must clearly correspond to the original photo at a glance. The bottom half uses childlike crayon, colored pencil, and marker doodles on off-white textured paper: crooked thick black outlines, visible crayon grain, repeated strokes, uneven coloring, occasional overshooting, naive perspective, and innocent proportions, like a city a child carefully drew from travel memories. Colors are extracted from the original photo and boosted with playful saturation, but different photos should not always use the same fixed color palette. The top and bottom each take about half, with only a clean thin white horizontal line in between. The bottom half must not trace the full photo in proportion; it should condense, extract, or reorganize the people and the most meaningful city details, while keeping breathing room from the off-white paper. The bottom half must be truly redrawn, not a photo filter. Do not invent extra landmarks, vehicles, people, or animals. Do not add titles, subtitles, labels, logos, watermarks, original image corner marks, AI corner marks, garbled text, nine-grid layouts, or multi-image collages; avoid refined vector illustrations, 3D rendering, and a realistic bottom half.