Collage image using ChatGPT

AI – 콜라주 프롬프트

여러 각도 사진 찍은 콜라주 만들기

​요즘 AI 공부를 하면서, Prompt Engineering도 같이 좀 보면서 놀고 있는데 사진을 다양한 각도로 보이게 좀 할 일이 있어서 테스트해봤다.

내 얼굴로 한 것은 좀 올리기 그래서, 나의 가상 친구 Lisa 으로 테스트한 결과.

순서대로 ChatGPT, Grok, Gemini의 결과이다.

어느 것이 더 자연스러운가?

​본인 사진 가지고 해 보면 재밌을 듯.

​난 본판이 엉망이라 AI를 써도 엉망으로 나온다.

Lisa face
나의 가상 친구 Lisa
Collage image by ChatGPT using Lisa face
Collage image by ChatGPT
Collage image by Grok using Lisa face
Collage image by Grok
Collage image by Gemini using Lisa face
Collage image by Gemini

테스트한 프롬프트

​아래 프롬프트를 그대로 복사해서, 본인 사진과 함께 ChatGPT등에 붙여넣기 하면된다.

그리고 실행하면 끝!

Use the uploaded face photo as the strict identity reference. Preserve facial structure, eye shape, nose, lips, skin tone, age, and overall likeness exactly. 

Do NOT beautify, stylize, cartoonize, or exaggerate facial features. The result must look like the same real person in real life. Photorealistic, natural human skin texture, real lighting. 

Create a 3x3 photo collage (9 images total) of the same person based on the uploaded face photo. 

STYLE: Ultra-photorealistic, real-life photography. Looks like DSLR camera photos, not AI-generated. Natural skin texture, pores, fine wrinkles, realistic shadows. No filters, no beauty retouching, no stylization. 

EXPRESSION: All images show a natural, friendly smile. Not exaggerated, not forced. Slight variations in smile intensity. 

ANGLES (one per image): Front-facing, neutral head position Slight left angle (15 degrees) Slight right angle (15 degrees) Left profile (45 degrees) Right profile (45 degrees) Slightly looking up Slightly looking down Head tilted slightly left while smiling Head tilted slightly right while smiling LIGHTING: Soft natural daylight, realistic indoor lighting. Consistent lighting across all images. No dramatic shadows, no studio effects. 

BACKGROUND: Simple, neutral background (light gray or off-white). No background distractions. CLOTHING: Same clothing in all images. Simple casual outfit, solid color. No logos or patterns. COMPOSITION: All images framed from shoulders up. Consistent camera distance. Arrange the final output as a clean 3x3 grid collage. 

IMPORTANT: The person must look exactly like the uploaded face photo. No changes to age, gender, ethnicity, or facial proportions. 

Avoid: cartoon, illustration, anime, CGI, painting, stylized art, beauty filter, face slimming, eye enlargement, plastic skin, overly smooth skin, unrealistic lighting, AI artifacts, different person, face drift, inconsistent identity

각 AI 서비스 바로가기

  • ChatGPT 바로가기

https://chatgpt.com

  • ​Grok 바로가기

https://grok.com

  • ​Gemini 바로가기

https://gemini.google.com

Leave a Reply

Your email address will not be published. Required fields are marked *