Over 10 years we help companies reach their financial and branding goals. Engitech is a values-driven technology agency dedicated.

Gallery

Contacts

411 University St, Seattle, USA

engitech@oceanthemes.net

+1 -800-456-478-23

Twitter Facebook-f Pinterest-p Instagram

Wrinkle Analysis across Time for Product Effectiveness

Category:
AI / FMCG Beauty Care

The Problem Statement

For the comprehensive analysis of skincare evaluations and the effectiveness of the skin and beauty creams, there is a need to detect and analyse skin issues like wrinkles present on the face across time. This can be achieved with a robust automated solution capable of precisely aligning face images taken across different timelines, after which we can use algorithms to detect, analyse and compare skin issues like wrinkles over time, giving information on the effectiveness of the skin care product.    

The Solution

We introduce an AI-powered solution that aligns facial images across different timestamps and quantifies wrinkle progression with high precision. Our advanced AI model processes facial images taken at different times, aligns them accurately, and performs an in-depth wrinkle analysis. By using deep learning techniques and image processing, the system detects fine lines and wrinkles across different skin tones and different age groups, providing an objective measure of facial aging.  

By aligning faces accurately across different timestamps and quantifying wrinkle progression, this innovation empowers skincare professionals, cosmetic brands, and individuals to make informed decisions about anti-aging treatments and skincare products. 

The App enables: 

  1. Accurate Face Alignment: Ensures consistency across different timestamps for reliable comparison. 
  2. Automated Wrinkle Detection: Identifies and analyse fine lines, deep wrinkles, and skin texture changes. 
  3. Quantification Metrics: Provides numerical data on wrinkle length, area, contrast, count, and spread (thickness) over time. 
  4. User-Friendly Interface: Simple and intuitive application for dermatologists, cosmetic companies, and individuals. 

Features of the App

Face alignment

  • Leveraging state-of-the-art computer vision techniques combined with advanced facial landmark detection algorithms, our system precisely identifies and matches facial features across images, irrespective of variations in pose, expression, or lighting conditions.
  • Pre-processing of the input image for converting from one colour space to others.
  • Developed an algorithm for detecting precise facial landmark points on frontal and profile views of high-resolution facial image captured by a clinical device.
  • Utilized a triangulation approach to align smaller regions of the face across different timestamps for the same subject.

Wrinkle detection

  • Manual annotation of wrinkles present in the high-resolution facial image data captured by the clinical device.
  • Developed a wrinkle-growing algorithm that generates a wrinkle mask from an annotated wrinkle, as per the appearance of the wrinkle, that is final wrinkle mask has variation in thickness across the length of the wrinkle.
  • AI model is trained on the wrinkle data, and further quantification of each of the wrinkles is done.
  • Quantification of wrinkles includes wrinkle count, length, area, thickness, and contrast.
  • This wrinkle detection algorithm can classify the detected wrinkle into fully developed (that is, thick and high contrast) and fine line (that is, thin and less contrast). It detects and categorizes wrinkles based on depth, length, and intensity.
  • Tracks changes over time with a detailed progression report.
  • Provides numerical scores for wrinkle severity and improvement over time.
  • This wrinkle detection algorithm is robust enough to work on different skin tones and across different age groups.

Visualization

The app interface features a simple upload process where users can submit images taken at different times. Our app automatically aligns these images, runs the wrinkle analysis, and presents the findings in an easy-to-understand format. A dashboard displays interactive reports with visual heatmaps and trend graphs showing wrinkle progression over time before and after using skin care product. 

Optimization and porting

We have optimized the face landmark detection module to efficiently detect face landmarks in the images and ported the Python code to C++ for efficiently aligning the different timestamp images for the app. We have also optimized the wrinkle detection AI model to run on the CPU to process the images faster for fine line and fully developed wrinkle detection. All the vision modules within the wrinkle detection algorithm were optimized along with the use of multi-threading for faster processing of the input images.  

Our app is robust enough to be integrated with skincare brands and clinics, enabling dermatologists and cosmetic companies to provide AI-driven wrinkle analysis to their customers. 

Benefits Delivered

  • In-depth insights into wrinkle analysis and personalized skincare recommendations.
  • Tracking the progression of wrinkles and the countereffect of recommended products over time.