概念留白剪影风上下对比转绘

Create a vertical 3:4 top-and-bottom comparative transfer drawing, style named [CONCEPTUAL SILHOUETTE VOID]. Applicable to people, animals, plants, architecture, cities, vehicles, food, vessels, interiors, nature, and travel photos, not limited to "journeys" or any single scene. 

The top half (about 42%–48%) retains the input photo itself: accurately maintain the main subject, quantity, pose, perspective, spatial relationship, light, material, and recognizable colors, with only natural publishing-level color grading, no illustration, no location change, and no adding or deleting key subjects. Leave a clean warm white seam between the top and bottom, without thick borders, photo tags, or Original/AI badges.

For the bottom half, do not copy the entire photo, nor draw a complete illustration. First, extract the most meaningful semantic relationship, action, silhouette, or material rhythm from the original photo: for example, a person's standing posture, the curve of an animal's tail, the opening of a flower, the extension of a tree branch, the bending of a road, the arching line of a building, the direction of a vehicle's speed, the circular relationship of dishes, the repeating rhythm of windows, the undulation of a mountain ridge, or the ripples on the water surface. Select only one main carrier and reconstruct it as a charcoal black, ink black, or dark graphite abstract flat silhouette, ribbon-like outline, blocky symbol, or single linear gesture. This shape must be traceable to the real subject in the top half, but allowed to be elongated, shrunk, offset, rotated, cut, or span the negative space.

The bottom half maintains 60%–80% continuous warm white negative space, with only very light fibers and matte grains on the paper. The edges of the main shape can have slight halftones, dry brush grains, or uneven ink density, but the overall style must be restrained, flat, and clean, not smudging into ink wash, nor becoming fine line art. Figures and animals can be compressed into overall silhouettes without faces or clothing textures; architecture, plants, food, and vessels retain only the minimum structure to identify the shape relationship, without restoring photo details. At most, use one or two extremely fine auxiliary lines to suggest direction; no second abstract object may be added.

The abstract shape is mainly charcoal black, ink black, or dark graphite, with at most one spot of extremely low-saturation color extracted from the original image; the color changes with the photo and does not use fixed blue, green, or city palettes. A row of very small, accurate English editorial notes can be added at the bottom to describe the semantics, such as ONE GESTURE / OPEN SPACE, ARC / STILLNESS, or SINGLE FORM; do not add large titles, long paragraphs, or fictional location information.

The overall design should let the top half bear the photographic facts and the bottom half bear the single concept, form judgment, and rhythm of negative space. Prohibit full-page scene redrawing, multiple floating objects, continuous route nodes, complete city skyline templates, photo windows, full-page ink wash, colored geometric posters, heavy paper art, 3D rendering, automatic outlining, logos, watermarks, fake Chinese, AI badges, and original image badges.

制作一张竖向 3:4 的上下对比转绘图，风格命名为【CONCEPTUAL SILHOUETTE VOID｜概念留白剪影风】。适用于人物、动物、植物、建筑、城市、交通工具、食物、器物、室内、自然和旅行照片，不限定“旅途”或任何单一场景。
上半部约 42%–48% 保留输入照片本身：准确保持主要主体、数量、姿态、视角、空间关系、光线、材质和可识别色彩，只做自然的出版级调色，不插画化、不换地点、不增删关键主体。上下之间留出干净的暖白接缝，不加粗边框、照片标签或原图/AI角标。
下半部不要复制整幅照片，也不要画完整插画。先从原照片提炼一个最有意义的语义关系、动作、轮廓或材质节奏：例如人物的站姿、动物的尾巴弧线、花的开放、树枝的伸展、道路的弯折、建筑的拱线、车辆的速度方向、碗碟的圆形关系、窗户的重复节奏、山脊的起伏或水面的波纹。只选择一个主载体，把它重构成一枚炭黑、墨黑或深石墨色的抽象平面剪影、带状轮廓、块面符号或单一线性手势。该形状必须能回溯到上半部的真实主体，但允许拉长、缩小、偏置、旋转、切断或跨越留白。
下半部保持 60%–80% 的连续暖白留白，纸面只有极淡纤维和哑光颗粒。主形状边缘可以有轻微网点、干刷颗粒或不均匀墨密度，但整体必须克制、平面、干净，不能晕成水墨，也不能变成精细线稿。人物和动物可压缩为无脸、无服装纹理的整体剪影；建筑、植物、食物和器物只保留能识别形状关系的最少结构，不恢复照片细节。最多使用一条或两条极细辅助线提示方向，不得添加第二个抽象物。
抽象形状以炭黑、墨黑或深石墨为主，最多从原图提取一处极低饱和点色，颜色随照片变化，不套固定蓝色、绿色或城市色板。底部可加入一行很小、准确的英文编辑注释，描述该语义，例如 ONE GESTURE / OPEN SPACE、ARC / STILLNESS 或 SINGLE FORM；不要加入大标题、长段落或虚构地点信息。
整体应让上半部承担摄影事实，下半部承担单一概念、形体判断和留白节奏。禁止满版场景重绘、多个漂浮物、连续路线节点、完整城市天际线模板、照片窗、整页水墨、彩色几何海报、厚重纸艺、3D渲染、自动描边、Logo、水印、假中文、AI角标和原图角标。