How DDD Enables an AI‑Driven Skin Analysis System for Customer Success

Project Background

Personalized skin‑care has become a major trend, prompting cosmetics brands to adopt AI skin‑measurement technology. An AI‑intelligent skin‑measurement system was proposed to provide an online skin‑management solution.

Users upload a mobile‑phone photo and receive a skin‑type report within 30 seconds. The AI analysis evaluates more than 30 skin dimensions, including age, dryness, oiliness, combination, sensitivity, blackheads, acne, spots, wrinkles, and aging. The system also tracks personal skin health, supports real‑time AI diagnosis, and offers personalized care recommendations.

DDD Business Modeling

During the early phase, the health‑management system was designed using the DDD four‑color modeling method. The health service was split into two aggregates:

File Management – manages C‑end customer basic data, analogous to an electronic medical record.

Health Management – handles skin‑analysis reports, modeled as a detection‑report entity.

The detection report is the core of the health‑management aggregate. Structured fields store common data; unstructured data such as report results and detection metrics are stored in JSON fields. Front‑end parsing is performed according to the report type, enabling flexibility and extensibility.

At present the report model supports dozens of types, including traditional Chinese‑medicine constitutions, blood tests, imaging, tongue diagnosis, pulse diagnosis, meridians, voice analysis, BMI, body‑composition analysis, blood pressure & heart rate, and psychological assessments, with plans to add more.

The underlying skin‑analysis API integrates Face++ capabilities, supplemented with custom scoring rules and product‑label matching algorithms to generate AI‑driven product recommendations.

Client Launch and Operations

The AI‑intelligent skin‑measurement system was developed intermittently over two months and delivered on schedule. The client opened a physical AI‑digital experience hall, installing robotic arms for photo capture and skin analysis. Daily foot traffic reached over 3,000 customers, distributors, and franchisees.

By the time of writing, the mini‑program had accumulated more than 80,000 registered users, over 120,000 skin‑profile records, and more than 9,000 orders, linking online and offline channels to sell AI‑customized skincare products nationwide.

Conclusion

The project demonstrates how DDD modeling combined with core product functions can create a tailored intelligent business solution. The four‑layer microservice architecture used for this project has been applied to multiple projects.

Source code for the DDD rapid‑start framework (D3Boot) is available at: https://gitee.com/jensvn/d3boot