// ==========================================
// AI Sims Filter - Backend Prompt Template
// ========================================== // [Option 1: Natural Language Prompt (Best for DALL-E 3 / Midjourney / Advanced LLM-based Image Models)]
Transform the person in the uploaded image into a 3D character from the video game "The Sims 4". Maintain the exact facial features, gender, and basic clothing shape of the original person, but convert them into a 3D stylized game asset. The skin should be smooth and slightly plastic or clay-like, without hyper-realistic pores or wrinkles. The hair should be styled in thick, stylized clumps ("Maxis Match" style), not individual strands. Crucially, add a glowing, translucent green diamond (Plumbob) floating directly above the character's head. The lighting should be bright, vibrant, and soft, resembling a high-quality 3D game engine render. Background should be a clean, solid pastel color or a stylized simple room. // [Option 2: Tag-based Positive Prompt (Best for Stable Diffusion / WebUI / API-based img2img)]
(The Sims 4 style:1.3), 3D character render, Maxis match aesthetic, game character avatar, floating glowing green Plumbob diamond above head, (smooth plastic-like skin:1.2), stylized thick hair, vibrant colors, EA game engine style, casual outfit, keeping original facial features, high quality 3D modeling, octane render, soft studio lighting, clean background, masterpiece, best quality, ultra-detailed. // [Negative Prompt (Crucial to prevent it from looking too realistic or creepy)]
photorealistic, realistic skin texture, visible pores, wrinkles, blemishes, messy hair, 2D anime, flat painting, sketch, dark lighting, horror, deformed, bad anatomy, ugly, (missing Plumbob:1.5), realistic eyes, text, watermark. // ==========================================
// Parameter Suggestions (If your API supports them):
// Image-to-Image Denoising Strength: 0.45 - 0.60 // (Keep it around 0.5: too low and it won't look like Sims, too high and it will lose the user's likeness)
// CFG Scale: 7.0 - 9.0
// ==========================================
请先填写必填项：Option 1: Natural Language Prompt (Best for DALL-E 3 / Midjourney / Advanced LLM-based Image Models), Option 2: Tag-based Positive Prompt (Best for Stable Diffusion / WebUI / API-based img2img), Negative Prompt (Crucial to prevent it from looking too realistic or creepy)
