Nano Banana & Nano Banana Pro Image Generation Guide

The awesome-nano-banana-images repository serves as a curated collection of diverse image generation and manipulation use cases. The project showcases a wide range of capabilities, from character design and 3D modeling to specialized tasks like infographic generation, product packaging, and architectural visualization.

Users can explore these cases to find inspiration or technical patterns for specific AI-driven image workflows. The cases are categorized into several types, including:

• Character & Persona: Custom character stickers, anime-to-real cosplayer conversion, character pose modification, and Pixar-style portraits.
• Design & Product: Product packaging generation, jewelry collection design, merchandise design, and logo typography.
• Spatial & 3D: Floor plan 3D renders, isometric building extraction, and generating ground views from map arrows.
• Utility & Editing: Automatic photo editing, image outpainting repair, watermark addition, and old photo colorization.
• Creative Art: Manga style conversion, comic book creation, and line art to doodle drawing.

Each case is attributed to a specific creator (e.g., @ZHO_ZHO_ZHO, @op7418, @icreatelife) and provides a specific functional example of what can be achieved with advanced image prompting and manipulation techniques.

The project follows specific rules for projecting source objects and rendering holographic outputs. Use these rules to guide your prompt engineering and setup.

Projection Source Rules

• 3D Object Input: Place a desktop 3D scanner next to the projector. Place the reference object on the scan plate. The hologram is generated from the scanned object.
• 2D Image Input: Place a modern PC with a monitor on a desk. Display the reference image on the monitor. The hologram is generated from the monitor content.

Hologram Rendering Rules

• Appearance: Characters are always displayed as translucent volume images with a faint background visible. No light rays, no particles, and no solid surface fragments.
• Anatomy & Silhouette: Maintain balanced anatomy (head-to-body ratio of 1/7 to 1/8) and clear silhouettes with natural poses.
• Details: Hair, clothing folds, and accessories must be visible but translucent. Faces must be sharp and expressive, legible even at a 1000px crop.
• Restrictions: No copyrighted characters, brand designs, or IP markings.

Environment & Camera Settings

• Environment: Modern office desk with a projector base and auxiliary equipment (scanner or monitor). Background is a seamless black studio with subtle reflections.
• Camera: 85-100mm lens, 3/4 hero angle, eye level, f/11~f/16, ISO100, tripod.
• Lighting: Soft lighting on the desk; holographic figures are defined by volume light only.

Output Specifications

• Format: 4:5 aspect ratio, 2048×2560 resolution.
• Sampling: Deterministic, Seed=12345, Temperature=0.
• Negative Prompting: Exclude text, watermarks, logos, brands, copyrighted characters, series IP, trademark designs, resin, PVC, solid statues, opaque surfaces, toy-like gloss, light rays, scan lines, dots, distortion, or unnecessary numbers.

To achieve specific results in Nano Banana, follow these prompting patterns observed in the examples:

1. Placeholder Substitution: For templates, replace text within [square brackets] with your specific details.

   • Example (Era/Style): 将角色的风格改为[1970]年代的经典[男性]风格 (Change character style to [1970]s classic [male] style).
   • Example (Object Type): 将图像制作成白天和等距视图[仅限建筑] (Make image into daytime and isometric view [architecture only]).

2. POI Specification: When using the AR experience generator, you must explicitly name the Point of Interest (POI) in the prompt.

   • Pattern: ...在这张图像中突出显示[兴趣点]并标注相关信息 (...highlight [POI] in this image and label relevant information).

3. Multi-Reference Coordination: When uploading multiple images, the prompt must explicitly describe the relationship between the references and the desired output.

   • Pattern: 准确使用图2色卡为图1人物上色 (Accurately use the color card in Image 2 to color the character in Image 1).

4. Detailed Scene Composition: For complex scenes involving multiple objects, provide a highly detailed description of every element, including colors, positions, and interactions.

This section provides various use cases for the Nano Banana image generation system, demonstrating how different input types (single images, multiple images, sketches, or text) combined with specific prompts can achieve diverse creative results.

Key capabilities demonstrated include:

• Character Transformation: Converting illustrations into physical figurines or realistic cosplayers.
• Spatial/Perspective Manipulation: Generating ground-level views from map arrows, creating top-down views from ground photos, or generating isometric models from standard images.
• Augmented Reality (AR) Simulation: Highlighting Points of Interest (POI) on real-world images.
• Style Transfer & Character Design: Changing eras (e.g., 1970s style), creating character design sheets (proportions, three-view drawings, expressions), or applying color palettes from a color card to line art.
• Image Enhancement & Editing: Automatic photo retouching (contrast, lighting, composition) and creating custom stickers.
• Multi-Reference Generation: Using multiple reference images to compose a complex scene with specific objects and characters.

<!-- Examples include: -->
// Example 1: Illustration to Figurine
// Prompt: "将这张照片变成角色手办..."

// Example 2: Map Arrow to Ground View
// Prompt: "从红色圆圈沿箭头方向画出真实世界的视角"

// Example 12: Character Design Sheet
// Prompt: "为我生成人物的角色设定 (Character Design)... 三视图 (正面、侧面、背面)..."

Nano Banana Pro is used for advanced creative workflows and complex image generation tasks. Examples include:

• Educational & Informational: Creating Ukiyo-e flashcards, generating flowcharts from documents, creating children's literacy newsletters, and generating biographies from Wikipedia.
• Character & Asset Design: Character cloning, character breakdown diagrams, character correlation charts, level evolution visuals, and creating crossover manga.
• Product & Industrial Design: Behind-the-scenes item production, material texture generation, toy disassembly displays, and creating product packaging.
• Artistic Styles: Generating images in the style of Qingming Shanghe Tu, creating movie storyboards, chalkboard chalk drawings, and isometric views.
• Data & Utility: Generating images from coordinates, creating PPTs from articles, estimating age from faces, and creating stylish maps.