AI Face Skin Analysis Model Development

Job ID: 39918444

Budget: ₹15,000 – ₹20,000 INR

Project Overview
Looking for an experienced AI/ML developer with strong skills in computer vision and deep learning to build an AI-powered face skin analyzer.
The goal is to allow users to scan or upload their face image, send it to the model for analysis, and receive a detailed skin report highlighting key skin concerns and overall skin type.

Reference:
The developed model and interface should provide similar capabilities, user experience, and analytical depth.
https://www.makeupar.com/business/showcase/skincare/home

Workflow
1. User scans or uploads their face image.
2. The image is sent to the backend AI model.
3. The model analyses the image and detects visible skin concerns.
4. Results are returned with details and coordinates to display visually on the user interface

Requirements / Model Output
The AI model should detect and return the following:
1. Acne Detection
* Detect acne regions with bounding box or segmentation coordinates
* Provide severity percentage
2. Redness Detection
* Identify if redness is present or not
* Return coordinates of affected areas (if any)
3. Wrinkle Detection
* Identify if wrinkles are present or not
* Return coordinates of affected areas (if any)
4. Dark Circle Detection
* Detect whether dark circles are present or not
5. Skin Type Classification
* Classify overall skin type: Oily / Dry / Normal / Combination / Sensitive
* Provide oiliness percentage and hydration percentage

Deliverables
* Fully trained and tested AI/ML model for all listed skin concern detections
* Complete codebase, including:
* Preprocessing scripts
* Model training and evaluation
* Inference scripts for running predictions

Timeline:
* Expected completion: 4-5 days

Ideal Candidate
* Strong background in computer vision, deep learning, and image segmentation/classification/detection
* Solid understanding of data preprocessing, augmentation, and model evaluation techniques
* Ability to deliver a clean, well-structured, and well-documented codebase