A beginner-friendly course on turning rough ideas into clear visual tasks, choosing the right tools, writing prompts, controlling composition and references, editing results, troubleshooting failures, and preparing images for safe delivery.
Choose between text-to-image, image-to-image, inpainting, outpainting, and layered editing based on the available assets, protected content, and permitted scope of change, then explain how text, images, masks, settings, and human judgment work together to control the result.
Choose a primary AI image workflow for a specific task based on starting assets, control requirements, revision methods, delivery requirements, and total effort, then explain which finishing tools should complement it.
Explain generation differences through model capability, current conditions, random sampling, and Seed, then choose a reasonable control action for excessive variation, ignored instructions, composition drift, and similar failures.
Turn a vague image request into a Visual Brief that covers purpose, platform, audience, core message, subject, visual direction, fixed and flexible content, and output limits, while separating generation, finishing, and acceptance information.
Extract the information needed for the current generation from a Visual Brief, then write a clear, editable prompt using subject, action, scene, composition, camera, lighting and color, style and materials, and only the necessary constraints.
Choose subject hierarchy, composition, shot size, camera angle, lighting, color, and materials for an image's purpose, then translate an abstract mood into visible conditions that reinforce one another.
Compare candidates through exploration, selection, convergence, and finalization; record protected elements, problems, and one change for the next round; and decide whether to continue, branch, or restart.
Choose references for content, composition, pose, style, color, people, products, or structure, then define each image's purpose, protected elements, flexible elements, and exclusions.
Separate fixed and flexible variables for a character, person, product, or brand series; build a reusable consistency card; and choose corrections based on the level of drift.
Distinguish content problems from setting problems, then choose canvas ratio, Seed, randomness, style strength, reference weight, variation strength, quality, and output settings for exploration, convergence, and finalization.
Choose regeneration, variation, inpainting, outpainting, or layered canvas editing for whole-image, directional, local, boundary, and compositing problems, then define masks, protected regions, and edit order.
Describe image problems as observable facts, locate them at the content, visual, technical, or delivery layer, and choose a cost-effective correction and stopping condition across prompts, references, settings, and editing methods.
Evaluate AI image risk across inputs, tools and models, generation, outputs, and publication; distinguish tool permission, copyright protection, and commercial responsibility; and plan disclosure, recordkeeping, and professional review.
Connect the Visual Brief, generation boundaries, tool workflow, prompt, references and settings, candidate selection, editing, quality review, risk review, and delivery records into an explainable AI image decision chain with clear rollback points.
