Our portfolios are dynamic records that show not only our completed projects but also the changing methods styles and visual languages we use as designers and architects. However as visualization trends change—from stylized schematics to photorealistic renderings—our older images may begin to seem out of place. With the Portfolio Enhancer a proof-of-concept interface that uses generative AI to update and harmonize your architectural visuals you can revitalize previous work harmonize styles throughout your collection and meticulously edit each image.
- Style Drift: Over time, software upgrades and emerging rendering paradigms can leave early projects looking dated.
- Visual Consistency: Mixed lighting setups, conflicting color palettes, and varying levels of detail can fracture a portfolio’s coherence.
- Creative Control: Off-the-shelf AI image tools often lack the granularity designers need to preserve a building’s form while reimagining its look.
The Portfolio Enhancer addresses these challenges head-on, offering architects a tailored workflow that keeps geometry intact while allowing full control over texture, lighting, and more.
How It Works
Import & Depth-Map Editing
Users import any existing project image when they first launch the application. A depth map which is a grayscale depiction of the scenes distances is produced by a first pass. Painting directly on this map allows you to specify which areas of the image the AI can alter. If you want to update the sky and landscaping while maintaining your meticulously modeled façade cover the building and let the rest change.
Prompt-Driven Refinement
Once you have locked in your geometry navigate to the prompt editor. AI can be guided toward your objective by a dropdown of pre-written prompts such as introducing realistic materials adjusting lighting or populating scenes with people that are scaled appropriately without causing any distortion to the built form. You can make sure that the AIs inventiveness enhances rather than replaces your design intent by choosing a targeted prompt.
Style-Model Selection
Lastly pick from a variety of pre-trained Laurel models each of which embodies a unique visual style such as high contrast conceptual diagrams moody dusk shots or minimalist white-box renders. By alternating between these models you can quickly iterate and find the ideal look to bring your portfolio together.
We continuously investigate how state-of-the-art computation can improve architecture in the IAAC Master in Advanced Computational Design. The Portfolio Enhancer shows how to revitalize our previous work and make sure each project shines with a cohesive modern vision by combining depth-map editing carefully chosen prompts and style-specific AI models. Await the next version where generative imagination and design precision collide if youre ready to give your portfolio an AI-powered makeover.
