# 写真加工

Generate a 3×3 grid avatar showing the person in this image with nine different hairstyles. Use the same person, the same outfit, and the same background in every cell, varying only the hairstyle (e.g., short bob, long curls, ponytail, braid, undercut, shoulder-length straight, layered cut, pixie, half-up bun). Preserve the person's identity, eye shape, nose, jawline, eyebrow shape, lip color, makeup state, clothing color, clothing fit, and pose exactly across all nine cells. Output as a single 3×3 grid image at the original aspect ratio, photorealistic and natural-looking with sharp focus per cell. Do not modify the facial structure or invent details that conflict with realistic hair physics, and avoid cartoon or anime stylization.

Auto-enhance this image to make it more vivid and engaging. Boost the contrast, deepen the color saturation, improve the lighting balance, and gently sharpen the existing details. Preserve the person's identity, eye shape, nose, jawline, hair length, hair color, hairstyle, eyebrow shape, lip color, makeup state, clothing color, clothing fit, pose, and facial expression exactly as the source. Keep the original composition, but you may lightly crop framing edges if needed. Output at the original resolution, photorealistic and natural-looking. Do not over-saturate the colors, do not introduce film grain or HDR look, and avoid any cartoon, painterly, or anime stylization.

Remove the unrelated bystanders walking in the background of the uploaded photo while keeping the main subject in the foreground completely untouched. Reconstruct the background — a Kyoto temple courtyard with stone paving and wooden architecture — so the cleanup looks seamless and continuous. Preserve the original lighting, shadows on the ground, and the subject's pose and expression. Output at the original resolution, photorealistic, do not invent details that conflict with the existing scene. Do not regenerate the photo from scratch — only operate on the existing pixels and preserve every face, pose, and the temple architecture exactly as the source.

Swap the outfit of the person in this image to a tailored navy single-breasted business suit with a crisp white dress shirt and a thin burgundy tie. Preserve the original pose, body proportions, hairstyle, facial identity, and background completely. Match the existing lighting on the new fabric with realistic wrinkles and natural shadows at the lapels and cuffs. Output as a 4:5 portrait, photorealistic. Do not alter the facial features or invent accessories that are not requested. Lock the exact facial identity and skin tone of the source person — do not regenerate the face.

Change the expression of the person in the uploaded photo from neutral to a noticeably warm closed-mouth smile that is clearly visible to the viewer, with a definite upward curve at the corners of the lips and gentle warmth around the eyes. Preserve the person's identity, facial proportions, hairstyle, and lighting completely. Avoid teeth showing and do not invent details around the mouth. Keep skin texture realistic and natural-looking. Output at the original resolution.

Crop and restyle the uploaded photo into a LinkedIn-ready professional headshot. Replace the background with a softly defocused modern office featuring floor-to-ceiling windows and warm ambient lighting. Preserve the person's facial features, expression, and outfit, including the exact color, fabric, and fit of the original clothing. Add subtle catchlights in the eyes and gentle three-point studio lighting on the face. Output as a 1:1 square at 1024×1024, photorealistic, do not over-smooth the skin.

Replace the background of the uploaded photo with a softly blurred professional photography studio backdrop in neutral light gray (#D8D8D8), with subtle vignetting toward the corners. Keep the person's pose, facial expression, hair flyaways, and clothing exactly as in the original. Do not change the hairstyle, hair length, or hair color. Match the existing key-light direction and 5500K color temperature on the subject so the composite looks natural. Output as a 1:1 square at 1024×1024 pixels. Photorealistic and natural-looking, with no white halo around the subject and no color cast on the skin from the new background.

Colorize the uploaded vintage black-and-white photo with historically plausible tones for the 1950s era: warm sepia-leaning skin tones, muted clothing colors appropriate to the period, and a softly desaturated background. Preserve the original composition, grain, and facial details exactly. Do not sharpen aggressively or invent details that are not visible in the source. Output at the original resolution, natural-looking and photorealistic. Do not regenerate the photo from scratch — only operate on the existing pixels and preserve every face, pose, and grain exactly as the source.

Change the hairstyle of the person in the uploaded photo to a chin-length French bob with blunt-cut bangs that just touch the eyebrows. Keep the original hair color, facial identity, expression, clothing, and background unchanged. Lock the exact facial identity — same eye shape, same nose, same face proportions as the source person. Match the lighting and shadow direction of the existing photo so the new hair blends naturally around the ears and neckline. Photorealistic, do not modify the facial structure. Output at the original resolution.

Softly retouch the face of the person in this image: reduce visible blemishes and minor redness, even out skin tone, and gently smooth pores while keeping skin texture, freckles, moles, and fine lines intact. Do not change the facial structure, eye shape, or jawline. Do not add makeup, lipstick, eyeliner, or modify the eyebrow shape — keep the bare-face look. Maintain the original pose, hair, and clothing. Output at the original resolution, photorealistic, do not over-smooth the skin and no plastic-looking face.

Transform the uploaded photo into a compliant Japanese passport ID photo following ICAO standards: plain white background (#FFFFFF), head centered and facing directly forward, neutral closed-mouth expression, both ears and full forehead visible, eyes open and looking at the camera, no shadows on the face or background. Preserve the person's natural skin texture, facial identity, and the original clothing (shoulders and collar must be visible — do not remove or alter the shirt) without beautification. Crop to 35×45mm with the head measuring 32–36mm from chin to crown. Output at 600 DPI. Do not generate bare skin or partial nudity.

Remove the background from the uploaded photo and isolate the person cleanly along the hair edges and shoulders. Preserve the original skin tones, clothing colors, and facial details exactly as they are. Keep the cutout sharp focus and natural-looking around the hair edges. Output as a transparent PNG at the original resolution with anti-aliased edges and no white halo around the subject.