Image generation
Image generation is the computational creation of new images from learned models, input data, or specified conditions.
Spring AI in Action
Image generation is the production of images by an AI model from textual or other input specifications.
- How it is used
- Integrates text-to-image models into Spring applications through a provider-neutral service and REST endpoint.
- How deep it goes
- Explains provider selection, autoconfiguration, service results, dimensions, and endpoint usage, without detailing model internals.
- Examples
- Uses a Burger Battle controller, Cowboy and Sunrise burger prompts, generated URLs, and saved PNG output.
- What you can do after
- Build it from scratch with Spring AI by configuring a provider, injecting ImageModel, and serving generated URLs or image bytes.
Generative AI in Action
Image generation is the computational production of visual images from learned patterns or specified inputs.
- How it is used
- Deploys visual generation and editing across e-commerce, entertainment, healthcare, creative production, synthetic data, and generative design.
- How deep it goes
- Explains major architectures, VAE latent states, prompting, and editing techniques, while stopping short of deep mathematical derivation.
- Examples
- Uses fashion design, virtual clothes try-on, medical imaging, faces, art, style transfer, and manually edited logos as examples.
- What you can do after
- Apply it via Stable Diffusion GUIs or APIs, craft stronger prompts, and select inpainting, outpainting, or variation workflows.
Read how Generative AI in Action explains image generation →
Deep Learning with R, Third Edition
Image generation is the computational production of images from learned data distributions or specified inputs.
- How it is used
- Frames creative image synthesis through latent-space sampling, iterative denoising, text conditioning, and latent-space editing.
- How deep it goes
- Explains latent spaces, model families, denoising, conditioning, prompts, and artifacts, while stopping short of formal derivations.
- Examples
- Demonstrates random-noise denoising, Stable Diffusion prompts, negative prompts, diffusion-step comparisons, and latent-space interpolation.
- What you can do after
- Apply it via KerasHub to generate prompted images, steer attributes with negative prompts, and vary diffusion steps.
Read how Deep Learning with R, Third Edition explains image generation →
Deep Learning with PyTorch, Second Edition
The task of creating new visual content from learned patterns, mentioned here as a preceding example task.
- How it is used
- Frames image synthesis as generative modeling that produces new visual samples from learned image-data distributions.
- How deep it goes
- Progresses from architectural history to diffusion fundamentals, noise schedules, embeddings, training, sampling, and a from-scratch implementation.
- Examples
- Uses pixels, image noise, beta schedules, sinusoidal embeddings, DALL-E, and Stable Diffusion as concrete examples.
- What you can do after
- Build a basic diffusion model from scratch and understand how it generates images by reversing progressively added noise.
Read how Deep Learning with PyTorch, Second Edition explains image generation →
Build a Text-to-Image Generator (from Scratch)
Image synthesis is the computational generation of new images from learned representations or specified conditions.
- How it is used
- Uses learned denoising models to synthesize images from noise, labels, or natural-language prompts.
- How deep it goes
- Traces noise initialization, 50-step denoising, scheduler updates, interpolation, prompting, resolution, and memory trade-offs.
- Examples
- Demonstrates labeled digits, generated flowers, interpolated images, and prompt-driven astronauts, animals, cities, and dogs.
- What you can do after
- Apply diffusion models through PyTorch functions or the diffusers StableDiffusionPipeline to generate and compare images.
Read how Build a Text-to-Image Generator (from Scratch) explains image generation →
AI for Everyday IT
AI image generation is the production of visual content from prompts or other inputs by an artificial intelligence model.
- How it is used
- Supports creative visuals and practical IT artifacts such as diagrams, mockups, logos, and technical documentation.
- How deep it goes
- Contrasts creative and technical uses, gives a prompt example, and stops short of model internals.
- Examples
- Uses a relieved cloud logo, Mermaid and PlantUML diagrams, LaTeX equations, HTML/CSS mockups, and infrastructure maps.
- What you can do after
- Apply it via DALL-E or GPT-4o, then refine outputs or convert technical markup into required enterprise formats.
Related concepts
- variational autoencoderused by · A variational autoencoder is a generative neural network that learns a probabilistic latent representation of data and decodes samples from it to reconstruct or generate data.
- diffusion modeluses · A diffusion model is a generative model that learns to produce data by reversing a process that progressively adds noise to data.
- DALL-Eimplemented by · DALL-E is a family of OpenAI generative models that creates and edits images from natural-language descriptions and other conditioning inputs.
- latent spaceuses · A latent space is a learned representational space in which numerical coordinates encode information or features inferred from observed data.
- GPT-4opart of · GPT-4o is an OpenAI multimodal generative model in the GPT family that processes and generates text and other supported modalities.
- generative modelinguses · Generative modeling is a machine-learning approach that learns patterns or a data distribution to generate new samples resembling the data.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.