
Nano Banana Pro
A pink-haired girl in a dark dress in a café, refined with soft lighting and smooth lines
Prompt
Repaint and refine the image as a whole without changing the original composition, character poses, character designs, original lighting effects, or camera relationships. Preserve the 2.5D Japanese illustration style, but reconstruct the image to create a cleaner, more cohesive visual result. Globally refine the colors and lighting of the existing image so that the overall light-and-shadow relationships follow a single, consistent lighting logic. Use a soft, even key light; avoid layered shadows and broken patches of color caused by multiple light sources. Shadow edges should transition naturally, with no hard boundaries or blocky fills. Unify the way materials are rendered, keeping the color transitions across skin, hair, and clothing continuous and gradient-based. Reduce localized noise, accumulations of color patches, and discontinuous shadow areas. Avoid any mosaic-like appearance, dirty blotches, or cloud-like color patches. Keep the overall style within the range of high-quality 2.5D animated illustration, with a subtle sense of volume and soft spatial depth, but do not introduce realistic skin texture, pore details, or photographic sharpening. Retain a moderate hand-drawn feel and smooth tonal gradations. Reduce interference from localized high-frequency details. Avoid overly sharp edges and excessive texture enhancement so the image has a clean visual flow and a unified color rhythm. Keep the character outlines and hair strands continuous, smooth, and consistent in direction. Avoid edge jitter, breaks, or grainy, jagged artifacts. Linework should resemble stable, hand-drawn animation line art, not boundaries formed by localized sharpening or accumulated noise. Contour changes should transition smoothly, with no localized “jumps” or irregular fluctuations. Overall, reduce microcontrast and the tendency toward localized sharpening so edge transitions retain a soft, visually blended quality. Avoid knife-sharp boundaries or overly distinct separation between skin, hair, and clothing. Highlights and shadows should connect through soft gradients rather than strong dividing lines. The overall impression should be “clear but not sharp”: retain form legibility while suppressing pixel-level sharpening artifacts. The final output should be a clear, soft, high-quality 2.5D illustration free of color contamination between regions, with consistent lighting, continuous linework, and natural gradient transitions.
在不改变原始构图、人物姿态、角色设计、原光影效果与镜头关系的前提下,对图像进行整体重绘优化。保持2.5D日系插画风格,但将画面重构为更干净、更统一的视觉输出。 对现有图像进行颜色与光照的全局重整,使整体明暗关系遵循单一稳定的光照逻辑。采用柔和、均匀的主光源,避免多光源造成的分层阴影与断裂色块。阴影边缘自然过渡,不出现硬边界或块状填色。 统一材质表现方式,使皮肤、头发与服装的色彩过渡保持连续渐变结构,减少局部噪点、色块堆积与非连续阴影区域。避免任何马赛克感、脏斑感或云状色块结构。 整体风格控制在高质量2.5D动画插画区间,具有轻微体积感与柔和空间深度,但不引入写实皮肤纹理、毛孔细节或摄影级锐化效果。保留适度手绘感与平滑色阶。 降低局部高频细节干扰,避免边缘过锐与纹理过度强化,使画面呈现干净的视觉流动性与统一的色彩节奏。 使人物轮廓线与发丝结构保持连续、顺滑且一致的走向,避免边缘抖动、断裂或颗粒化锯齿感。线条表现应接近稳定的手绘动画线稿,而不是由局部锐化或噪声堆叠形成的边界。轮廓变化应平滑过渡,不出现局部“跳变”或不规则波动。 整体降低微观对比度与局部锐化倾向,使边缘过渡保持轻柔的视觉融合感。避免皮肤、头发与服装边界出现刀切式硬边或过度清晰的分离感。高光与阴影之间应以柔性渐变连接,而非强分界线式结构。整体观感为“清晰但不锐利”,保留形体辨识度 but 抑制像素级锐化痕迹。 最终输出应呈现为清晰、柔和、无分区污染的高质量2.5D插画画面,具有稳定的光照体系、连续的线条结构与自然的渐变过渡。
The rights to this prompt belong to its author @biantai0721. You are responsible for any commercial use — promoting AI tools, aggregators or Telegram bots: mozgi.io accepts no liability. Details →











