Contact person: Laravel AI Interior Element Segmentation

Phone:Show

Email:Show

Location: Saudi Arabia

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I need to plug a Python-based AI module into my existing Laravel site so visitors can upload interior photos and instantly receive separate PNG layers for every detected element. The model must recognise furniture, decorative items, architectural features, as well as floors, walls and ceilings, then export each class to its own transparent layer for further editing.

The flow I have in mind is simple: Laravel handles the authentication and file upload, hands the image to your Python service (Docker or local venv, whatever you prefer), and receives back a structured ZIP (or similar) containing one PNG per element plus a JSON manifest listing the class names and bounding masks. A lightweight REST or gRPC bridge will be enough—no need to over-engineer.

I’m open to the framework you choose—PyTorch, TensorFlow, Detectron2, Segment Anything, or a custom U-Net—so long as the final output is accurate on common residential scenes and easy for my team to extend with new classes later.

Deliverables
• Python segmentation module with trained weights
• Integration endpoint callable from Laravel (controller stub and example request)
• Clear README covering setup, retraining and class mapping
• A short test suite (5–10 sample images + expected layers) confirming everything works out of the box

Once it runs locally we’ll deploy it to our VPS, so keep dependencies lean and document any GPU needs. If this sounds straightforward to you, tell me how long you’ll need to train/tune the model and how you plan to evaluate accuracy." (client-provided description)