← back
AI Facial Analysis Mobile App

AI Facial Analysis Mobile App

Pending
💰 USD 30–250 👤 Unknown 🕒 4d ago status: new
Python Mobile App Development iPhone Android Machine Learning (ML) OpenCV iOS Development Computer Vision
My goal is to build a production-ready mobile application that ingests four user-supplied facial photos and returns a detailed, clinically oriented assessment. The core workflow will run on-device or securely in the cloud, extracting landmarks, analysing colour data, and generating a comprehensive report that includes: • Wrinkle and fine-line detection • Skin-texture metrics (smoothness, pore visibility, roughness) • Facial-symmetry scoring across key planes In addition, the engine should render colourimetric maps, redness overlays, and force-vector visualisations suitable for orofacial harmonisation planning. Target platforms The front-end must launch natively on iOS and Android with a shared codebase where practical. A lightweight web dashboard for clinicians is a welcome bonus but not required for the first milestone. Tech stack preferences I am open to leveraging OpenAI, Gemini, or Claude for model hosting or prompt-based post-processing, alongside tried-and-true computer-vision libraries such as OpenCV, Dlib, or MediaPipe for landmark detection and image preprocessing. If you favour a different modern framework, outline your rationale. Key expectations • Proven experience stitching together CV pipelines and ML inference for medical or aesthetic imaging • Clear, well-commented code and an API/SDK surface that lets me plug additional metrics in later • HIPAA/GDPR-conscious data handling; photos never persist longer than necessary • Regular English-language progress updates and demo builds Deliverables (first release) 1. iOS and Android apps that accept four images, run the analysis, and display results in an easy-to-read report view 2. Back-end or on-device model code with reproducible training/inference scripts 3. Developer-level documentation covering build, deployment, and extension points 4. Short validation dataset proving accuracy of wrinkle, skin texture, and symmetry modules Acceptance criteria A sample test face must generate consistent metric values within ±5 % across repeated runs, and overlay visualisations should align within ≤3 px of ground-truth landmarks on the validation set. If this brief matches your expertise, share a succinct plan of attack, relevant portfolio links, and an estimated timeline to reach the first working prototype.
↗ View on Freelancer