角色提示詞

Miss Nancy is an older African-American woman with pink hair rollers, a pink robe, pink slippers, large round glasses, and big expressive bug eyes. She has a nosy, dramatic personality and exaggerated facial expressions.

Scene takes place inside her living room during the daytime. The room is slightly messy with curtains half open, sunlight shining in, and a couch near the window.

Miss Nancy is standing very close to an Alexa speaker on a table, leaning in suspiciously. She whispers loudly, then suddenly yells, thinking Alexa is spying on her. Her bug eyes widen dramatically, and she clutches her robe.

She starts arguing with Alexa like it’s a real person, pacing back and forth. She points at it, gasps, then backs up slowly like she’s scared. Then she quickly grabs it, shakes it, and demands answers.

Background sounds: light TV static, birds chirping outside, faint neighbor noise through the wall.

Facial expressions: exaggerated, wide eyes, mouth dropping open, dramatic side-eyes, confused blinking.

Camera: medium close-up, slight zoom-in when she gets dramatic.

Lighting: bright daytime, soft shadows.

Style: colorful, cartoon, not realistic.

No text on screen. No subtitles. No watermarks.

Act as an Algorithm Analysis and Improvement Advisor. You are an expert in artificial intelligence and computer vision algorithms with extensive experience in evaluating and enhancing complex systems. Your task is to analyze the provided algorithm and offer constructive feedback and improvement suggestions.

You will:
- Thoroughly evaluate the algorithm for efficiency, accuracy, and scalability.
- Identify potential weaknesses or bottlenecks.
- Suggest improvements or optimizations that align with the latest advancements in AI and computer vision.

Rules:
- Ensure suggestions are practical and feasible.
- Provide detailed explanations for each recommendation.
- Include references to relevant research or best practices.

Variables:
- ${algorithmDescription} - A detailed description of the algorithm to analyze.

Act as an Algorithm Expert. You are an expert in algorithms with extensive experience in explaining and breaking down complex algorithmic concepts for learners of all levels.
Your task is to provide clear and concise explanations of various algorithms.
You will:
- Summarize the main idea of the algorithm.
- Explain the steps involved in the algorithm.
- Discuss the complexity and efficiency.
- Provide examples or visual aids if necessary.
Rules:
- Use simple language to ensure understanding.
- Avoid unnecessary jargon.
- Tailor explanations to the user's level of expertise (beginner, intermediate, advanced).
Variables:
- ${algorithmName} - The name of the algorithm to explain
- ${complexityLevel:beginner} - The level of complexity to tailor the explanation

Photorealistic iPhone selfie-style shot in alpine mountains. Bright clear daylight, deep blue sky, dramatic sharp mountain peaks in the background with patches of snow on rocky ridges. Wide open green alpine meadow in the foreground, lush grass with small plants visible in detail. A small wooden mountain hut in the mid-distance. The woman lies on her back in the grass, relaxed, using a hiking backpack as a pillow. The camera angle is handheld and slightly above her — classic iPhone arm-extended selfie perspective, subtle wide-angle distortion on the extended arm. She wears sporty hiking outfit: lightweight Arc’teryx windbreaker jacket (blue tone), fitted pink athletic shorts, Oakley sunglasses, casual trail vibe. Relaxed body posture — one knee slightly bent, one arm extended toward the camera holding the phone. Backpack visible under her head, realistic hiking gear details.

{
  "title": "Alpine Freefall",
  "description": "A high-octane, wide-angle action shot capturing the exhilarating rush of a freestyle skier mid-descent on a steep mountain peak.",
  "prompt": "You will perform an image edit using the person from the provided photo as the main subject. Preserve her core likeness. Create a hyper-realistic GoPro selfie-style image of Subject 1 speeding down a high-altitude ski slope. The image should feature the signature fisheye distortion, capturing the curvature of the horizon and the intense speed of the descent, with the subject holding the camera pole to frame herself against the dropping vertical drop.",
  "details": {
    "year": "2024",
    "genre": "GoPro",
    "location": "A jagged, snow-covered mountain ridge in the French Alps with a clear blue sky overhead.",
    "lighting": [
      "Bright, harsh sunlight",
      "Lens flare artifacts",
      "High contrast"
    ],
    "camera_angle": "Selfie-stick POV with wide-angle fisheye distortion.",
    "emotion": [
      "Exhilarated",
      "Fearless",
      "Wild"
    ],
    "color_palette": [
      "Blinding white",
      "Deep azure",
      "Stark black",
      "Skin tones"
    ],
    "atmosphere": [
      "Adrenaline-fueled",
      "Fast-paced",
      "Crisp",
      "Windy"
    ],
    "environmental_elements": "Kicked-up powder snow spraying towards the lens, motion blur on the edges, water droplets on the camera glass.",
    "subject1": {
      "costume": "black mini skirt, white crop top, leather fingerless gloves",
      "subject_expression": "Wide-mouthed shout of excitement, eyes wide with the thrill.",
      "subject_action": "ski"
    },
    "negative_prompt": {
      "exclude_visuals": [
        "studio lighting",
        "calm",
        "static pose",
        "indoor settings",
        "trees"
      ],
      "exclude_styles": [
        "oil painting",
        "sketch",
        "warm vintage",
        "soft focus"
      ],
      "exclude_colors": [
        "sepia",
        "muted tones",
        "pastel"
      ],
      "exclude_objects": [
        "ski lift",
        "crowd",
        "buildings"
      ]
    }
  }
}

Act as a Digital Inclusion Specialist focused on Web Accessibility (A11Y). Your sole mission is to generate high-quality alternative text (Alt Text) that provides visually impaired users with an equitable and vivid understanding of images through screen readers.

Follow these strict WCAG-aligned principles:
1. **Directness:** Never use "Image of" or "Photo of." Start describing the scene immediately.
2. **The 125-Character Rule:** Be concise. Convey the core meaning in about 125 characters. If the image is complex (e.g., an infographic), provide a concise summary of the key message.
3. **Hierarchy of Information:** Identify the primary subject first, then mention essential spatial relationships or background elements that define the context.
4. **Objective Description:** Describe what is physically visible. Avoid subjective interpretations (e.g., instead of "beautiful scenery," use "golden hour sunlight hitting a calm lake").
5. **Text Representation:** If the image contains text, transcribe it exactly within quotes.
6. **Atmosphere:** Briefly mention the mood or lighting if it's crucial to the visual's intent (e.g., "dimly lit," "high-contrast," "vibrant").

### Output Schema:
- **Alt Text:** [Place the descriptive text here]

### Few-Shot Examples:
- **Input:** [A photo of a guide dog leading a person across a busy city street]
- **Alt Text:** A golden retriever guide dog in a harness leads a person across a marked crosswalk on a busy city street with cars stopped.
- **Input:** [A minimalist digital flyer for a bake sale on Friday at 4 PM]
- **Alt Text:** Minimalist flyer with "Bake Sale" in bold font. Details: "Friday at 4 PM." Background features simple line drawings of cookies.
- **Input:** [A close-up of a person's hands knitting a blue wool scarf]
- **Alt Text:** Close-up of hands using wooden needles to knit a textured, bright blue wool scarf.

Now, analyze the provided image and generate the most inclusive Alt Text possible.

Amateur girls’ night selfie, very casual and imperfect, 1:1 aspect ratio. The image is shot directly from the FRONT CAMERA of a cheap, older smartphone: we see only what the phone sees, we DO NOT see any phones or cameras in the frame.

Three adult women sit close together on an old, comfy couch in a small apartment living room at night. They are wearing simple home clothes and sweatpants, like a real chill night in.

Center woman: medium skin tone, long dark hair, wearing a plain black sleeveless top and light grey sweatpants. She sits in the middle of the couch, one leg tucked under her, the other bent. Her body leans slightly toward the left, head tilted a bit, smiling softly toward the camera, relaxed and unposed.

Left woman: light skin and straight, light-brown hair, wearing a long-sleeve black top and light grey sweatpants. She leans in very close to the center woman, almost touching shoulders, making a big exaggerated kissy face toward the camera, lips puckered, eyebrows slightly raised. Because this is a selfie POV, she appears slightly closer and a bit larger from perspective, like someone near the phone.

Right woman: light skin and wavy blonde hair, wearing a dark long-sleeve top and black leggings. She leans into the group from the right, head tilted, smiling with her tongue out in a playful, goofy expression, eyes squinting slightly from laughter. All three look like close friends having fun, not models.

Environment: cozy, slightly messy living room. Behind them, a simple floor lamp with a warm bulb lights the wall. In the background on one side, a TV screen is visible with a paused movie scene (soft, abstract shapes, no recognizable faces or logos). On a low wooden coffee table in front of the couch (visible at the bottom of the frame) are open pizza boxes with half-eaten slices, a bag of chips, a soda can and a sparkling water can, a few crumbs, and a phone lying flat on the table. The room has string lights or fairy lights along one wall, giving a warm, imperfect glow. The apartment and furniture look normal and slightly worn, not like a studio set.

Camera and style: VERY IMPORTANT – this image should look like a real, bad selfie, NOT a professional photo. It is captured with a basic smartphone front camera in AUTO mode. Direct, slightly harsh phone flash from near the lens, with faces a little overexposed and shiny in some spots. Visible digital noise and grain in the darker parts of the room. Mixed lighting: warm yellow from the lamp and a cooler bluish cast from the TV, giving slightly uneven white balance. Focus is soft, not razor sharp, with a tiny bit of motion blur in hair and hands. Edges of the frame have mild vignetting and slight wide-angle distortion, like a cheap front camera. The composition is a little crooked and off-center; some pizza boxes and objects are cut off at the edges. Overall, the picture should feel like an unedited, spontaneous selfie sent to a group chat.

Constraints: there are EXACTLY THREE women in the frame and NO other people. The only camera is the phone we are looking through, so no extra hands, no extra phones, no mirror showing the photographer, no second photographer at the edge of the frame. No reflections of another camera. Just the three friends on the couch and the messy coffee table.

Negative prompt: professional studio, pro lighting, softboxes, rim light, cinematic atmosphere, commercial photoshoot, perfect color grading, HDR, strong depth of field blur, bokeh, high-end DSLR or lens, ultra-clean fashion image, symmetrical composition, influencer preset, heavy airbrushed skin, filters, hotel room, staged set, extra people, extra arms, extra hands, any additional phones or cameras in the frame, mirrors showing another photographer, text, logo, watermark, surreal glitches, underage appearance.

instagirl, mirror selfie in a hallway, realistic amateur phone snapshot, natural skin texture, minimal makeup, mild lens distortion from phone camera, casual posture, everyday outfit, slight handheld micro-blur, iPhone 11 wide 26mm EXIF feel, imperfect framing (a little headroom cut), mixed indoor lighting with slight color cast, background clutter present, no retouching, no beauty filter, faithful anatomy, same person identity, same body proportions, match reference face closely, iphone 11 pro max,

story: a child superman and a child batman joins their forces together in a forest. it's a beautiful day in the forest and they see a stick shelter and want to check out. they see a fox and for several seconds both fox and kids don't know what to do. they think first. then they all decide to run in opposite directions

instructions: {
  "style": {
    "name": "American Comic Book",
    "description": "Bold, dynamic comic book page in the classic American superhero tradition. Deliver your narrative as a fully realized comic page with dramatic panel layouts, cinematic action, and professional comic book rendering."
  },
  "visual_foundation": {
    "medium": {
      "type": "Professional American comic book art",
      "tradition": "DC/Marvel mainstream superhero comics",
      "era": "Modern age (2000s-present) with classic sensibilities",
      "finish": "Fully inked and digitally colored, publication-ready"
    },
    "page_presence": {
      "impact": "Each page should feel like a splash-worthy moment",
      "energy": "Kinetic, explosive, larger-than-life",
      "tone": "Epic and dramatic, never static or mundane"
    }
  },
  "panel_architecture": {
    "layout_philosophy": {
      "approach": "Dynamic asymmetrical grid with dramatic variation",
      "pacing": "Panel sizes reflect story beats—big moments get big panels",
      "flow": "Clear left-to-right, top-to-bottom reading path despite dynamic layout",
      "gutters": "Clean white gutters, consistent width, sharp panel borders"
    },
    "panel_variety": {
      "hero_panel": "Large central or full-width panel for key action moment",
      "establishing": "Wide panels for scale and environment",
      "reaction": "Smaller panels for faces, dialogue, tension beats",
      "inset": "Occasional overlapping panels for emphasis or simultaneity"
    },
    "border_treatment": {
      "standard": "Clean black rectangular borders",
      "action_breaks": "Panel borders may shatter or be broken by explosive action",
      "bleed": "Key moments may bleed to page edge for maximum impact"
    }
  },
  "artistic_rendering": {
    "line_work": {
      "quality": "Bold, confident, professional inking",
      "weight_variation": "Heavy outlines on figures, medium on details, fine for texture",
      "contour": "Strong silhouettes readable at any size",
      "hatching": "Strategic crosshatching for form and shadow, not overworked",
      "energy_lines": "Speed lines, impact bursts, motion trails for kinetic action"
    },
    "anatomy_and_figures": {
      "style": "Heroic idealized anatomy—powerful, dynamic, exaggerated",
      "musculature": "Detailed muscle definition, anatomy pushed for drama",
      "poses": "Extreme foreshortening, dramatic angles, impossible dynamism",
      "scale": "Figures commanding space, heroic proportions",
      "expression": "Intense, readable emotions even at distance"
    },
    "environmental_rendering": {
      "destruction": "Detailed rubble, debris clouds, structural damage",
      "atmosphere": "Rain, smoke, dust, particle effects for mood",
      "architecture": "Solid perspective, detailed enough for scale reference",
      "depth": "Clear foreground/midground/background separation"
    }
  },
  "color_philosophy": {
    "approach": {
      "style": "Modern digital coloring with painterly rendering",
      "depth": "Full modeling with highlights, midtones, shadows",
      "mood": "Color supports emotional tone of each panel"
    },
    "palette_dynamics": {
      "characters": "Bold, saturated colors for heroes/main figures",
      "environments": "More muted, atmospheric tones to push figures forward",
      "contrast": "Strong value contrast between subjects and backgrounds",
      "temperature": "Strategic warm/cool contrast for depth and drama"
    },
    "atmospheric_coloring": {
      "sky": "Dramatic gradients—stormy grays, apocalyptic oranges, moody blues",
      "weather": "Rain rendered as white/light blue streaks against darker values",
      "fire_energy": "Vibrant oranges, yellows with white-hot cores, proper glow falloff",
      "smoke_dust": "Layered opacity, warm and cool grays mixing"
    },
    "lighting_effects": {
      "key_light": "Strong dramatic source creating bold shadows",
      "rim_light": "Edge lighting separating figures from backgrounds",
      "energy_glow": "Bloom effects on power sources, eyes, weapons",
      "environmental": "Bounce light from fires, explosions, energy blasts"
    }
  },
  "typography_and_lettering": {
    "speech_bubbles": {
      "shape": "Classic oval/rounded rectangle balloons",
      "border": "Clean black outline, consistent weight",
      "tail": "Pointed tail clearly indicating speaker",
      "fill": "Pure white interior for maximum readability"
    },
    "dialogue_text": {
      "font": "Classic comic book lettering—bold, clean, uppercase",
      "size": "Readable at print size, consistent throughout",
      "emphasis": "Bold for stress, italics for whispers or thoughts"
    },
    "sound_effects": {
      "style": "Large, dynamic, integrated into the art",
      "design": "Custom lettering matching the sound—jagged for explosions, bold for impacts",
      "color": "Vibrant colors with outlines, shadows, or 3D effects",
      "placement": "Part of the composition, not just overlaid"
    },
    "captions": {
      "style": "Rectangular boxes with subtle color coding",
      "placement": "Top or bottom of panels, clear hierarchy"
    }
  },
  "action_and_dynamics": {
    "motion_rendering": {
      "speed_lines": "Radiating or parallel lines showing movement direction",
      "motion_blur": "Selective blur on fast-moving elements",
      "impact_frames": "Starburst patterns at point of collision",
      "debris_scatter": "Rocks, glass, rubble flying with clear trajectories"
    },
    "impact_visualization": {
      "collision": "Visible shockwaves, ground cracks, structural deformation",
      "energy_attacks": "Bright core fading to colored edges with atmospheric scatter",
      "physical_force": "Bodies reacting realistically to impossible forces"
    },
    "camera_dynamics": {
      "angles": "Extreme low angles for power, high angles for scale",
      "foreshortening": "Aggressive perspective on approaching figures/fists",
      "dutch_angles": "Tilted frames for tension and unease",
      "depth_of_field": "Suggested focus through detail level and blur"
    }
  },
  "atmospheric_elements": {
    "weather": {
      "rain": "Diagonal streaks, splashes on surfaces, wet reflections",
      "lightning": "Bright forks illuminating scenes dramatically",
      "wind": "Debris, hair, capes showing direction and force"
    },
    "destruction_aesthetic": {
      "rubble": "Detailed concrete chunks, rebar, shattered glass",
      "dust_clouds": "Billowing, layered, atmospheric perspective",
      "fire": "Realistic flame shapes with proper color temperature gradient",
      "smoke": "Rising columns, drifting wisps, obscuring backgrounds"
    },
    "scale_indicators": {
      "buildings": "Damaged structures showing massive scale",
      "vehicles": "Cars, tanks as size reference objects",
      "crowds": "Smaller figures emphasizing main subject scale"
    }
  },
  "technical_standards": {
    "composition": {
      "focal_point": "Clear visual hierarchy in every panel",
      "eye_flow": "Deliberate path through panels via placement and contrast",
      "balance": "Dynamic asymmetry that feels intentional, not chaotic"
    },
    "consistency": {
      "character_models": "Consistent design across all panels",
      "lighting_logic": "Light sources make sense across the page",
      "scale_relationships": "Size ratios maintained throughout"
    },
    "print_ready": {
      "resolution": "High resolution suitable for print reproduction",
      "color_space": "Vibrant colors that work in CMYK",
      "bleed_safe": "Important elements away from trim edges"
    }
  },
  "page_composition": {
    "no_border": {
      "edge_treatment": "NO frame around the page—panels extend to image edge",
      "bleed": "Page IS the comic page, not a picture of one",
      "presentation": "Direct comic page, not photographed or framed"
    }
  },
  "avoid": [
    "Any frame or border around the entire page",
    "Photograph-of-a-comic-page effect",
    "Static, stiff poses without energy",
    "Flat lighting without dramatic shadows",
    "Muddy, desaturated coloring",
    "Weak, scratchy, or inconsistent line work",
    "Confusing panel flow or layout",
    "Tiny unreadable lettering",
    "Sound effects as plain text overlay",
    "Anatomically incorrect figures (unless stylized intentionally)",
    "Empty, boring backgrounds",
    "Inconsistent character scale between panels",
    "Manga-style effects in American comic aesthetic",
    "Overly rendered to the point of losing graphic punch",
    "Weak impact moments—every action should have weight"
  ]
}

Kodak porra 400 Authentic vintage analog film photography, captured on classic 35mm film camera with manual focus lens, shot on expired Kodak Portra 400 film stock, pronounced natural film grain structure with visible halation around bright highlights, warm nostalgic color palette with slightly desaturated mid-tones, organic color shifts between frames, gentle peachy skin tones characteristic of Portra film, soft dreamy vignetting gradually darkening towards corners and edges, accidental light leaks with orange and red hues bleeding into frame edges, subtle lens flare from uncoated vintage optics, imperfect manual focus creating dreamy bokeh with swirly out-of-focus areas, chromatic aberration visible in high contrast edges, film dust particles and hair caught during scanning process, fine vertical scratches from film transport mechanism, authentic analog warmth with slightly lifted blacks and compressed highlights, natural color bleeding between adjacent film layers, gentle overexposure in bright areas creating soft glow, film edge artifacts and frame numbers barely visible, scanned from original negative with slight color cast, 1990s point-and-shoot disposable camera aesthetic, Fujifilm Superia or Agfa Vista alternative film characteristics, organic photographic imperfections and inconsistencies, slightly soft focus overall sharpness, date stamp in corner optional, double exposure ghost images subtle overlay, sprocket holes impression, cross-processed color shifts, pushed film development look with increased contrast and grain, natural lighting artifacts and lens imperfections, retro photo lab color correction style, authentic film emulsion texture, varying exposure between frames showing human photographer touch, mechanical shutter artifacts, slight motion blur from slower shutter speeds, nostalgic summer afternoon golden hour warmth, faded photograph found in old shoebox quality, memory lane aesthetic, tactile analog photography feel