AI 人像对比侧边展示

中文提示词

一张桌面侧边对比截图，在两个独立的 macOS 风格图像查看器窗口中展示两张几乎相同、由 AI 生成的街头人像，背景为浅青色。布局包含左右两个垂直图像面板，每个面板位于带有顶部工具栏的圆角灰色窗口内。左侧窗口标题以“NanoBanana2”开头，右侧窗口标题以“ChatGPTImage2”开头。两个面板均展示同一场景概念：一位时尚年轻女性站在黄金时段的城市街角，取景范围从大腿中部到头部，面部被柔和的矩形模糊块遮挡。她留着波浪形棕色长发，身材苗条，身穿露肩修身米色长袖上衣、高腰蓝色牛仔裤，佩戴精致的小吊坠项链，单肩背着一个带有金色链条的黑色大皮包。她的一只手拿着黑色智能手机，置于腰部附近。两个面板的背景均为高档绿树成荫的城市街道，温暖的阳光透过树叶洒下，具有浅景深效果，画面中有停放和行驶的车辆、古老的石砌建筑，每张图像左上角附近可见绿色的交通信号灯。保持两张人像非常相似但不完全相同：左侧图像色调稍冷、更柔和且更柔和，采用居中的街道视角；右侧图像色调稍暖、更明亮且更清晰，头发上的阳光感更强，虚化效果更具戏剧性。包含两个由窄垂直间隙隔开的查看器窗口，带有细腻的投影和简洁的科技对比展示效果，如同捕捉两个图像生成模型之间的视觉对比。

原始提示词

A side-by-side desktop screenshot comparing two nearly identical AI-generated street portraits in separate macOS-style image viewer windows against a light cyan background. The layout has exactly 2 vertical image panels, left and right, each inside a rounded gray window with a toolbar at the top. The left window title begins with "NanoBanana2" and the right window title begins with "ChatGPTImage2". In both panels, show the same scene concept: a fashionable young woman standing on a city street corner during golden hour, framed from about mid-thigh to head, with her face obscured by a soft rectangular blur block. She has long wavy brown hair, a slim figure, and wears an off-shoulder fitted cream long-sleeve top, high-waisted blue jeans, a delicate small pendant necklace, and a large black leather shoulder bag with a gold-tone chain strap over one shoulder. She holds a black smartphone in one hand near her waist. The background in both panels is an upscale tree-lined urban street with warm sunlight filtering through leaves, shallow depth of field, parked and passing cars, old stone buildings, and a visible green traffic light near the upper left area of each image. Keep the two portraits very similar but not identical: the left image is slightly cooler, softer, and more muted with a centered street perspective; the right image is slightly warmer, brighter, and crisper with stronger sunlight on the hair and more dramatic bokeh. Include exactly 2 viewer windows separated by a narrow vertical gap, with subtle drop shadows and a clean tech-comparison presentation, as if capturing a visual comparison between two image generation models.