-
-
output
-
data collecting and analysing in realtime
Inspiration
Beauty has always been seen as subjective—different cultures, eras, and people define it in their own ways. But what if technology could decode beauty mathematically, using the Golden Ratio and AI? That question sparked Beauty-Ratio. I wanted to see if a machine could interpret facial aesthetics not just as numbers, but as patterns of symmetry and proportion that humans subconsciously recognize as beautiful.
What it does
Beauty-Ratio is a real-time Python application that evaluates facial proportions, symmetry, and skin tone. It: Detects 3D facial landmarks. Compares facial ratios against the Golden Ratio (≈1.618). Predicts a beauty score using a neural network (Perfect, Good, Average, Bad). Analyzes skin tone using HSV color space. Provides live visual feedback on areas that align with or deviate from ideal symmetry. In short: it merges mathematics, AI, and art to quantify beauty in a way that feels both scientific and fascinating.
How we built it
We combined computer vision, deep learning, and geometry to bring this idea to life:
OpenCV → For real-time video capture and visualization. MediaPipe → To detect and map 3D facial landmarks with high precision. TensorFlow → To build and train a neural network that predicts beauty scores. NumPy → For numerical computations and Euclidean distance-based facial measurements. Golden Ratio Logic → Used as the benchmark to evaluate facial harmony. Every frame from the webcam is analyzed instantly, turning raw data into meaningful beauty insights.
Challenges we ran into
Real-time performance: Balancing neural network accuracy with smooth video processing. Data limitations: Training a beauty score predictor with limited datasets. Lighting issues: Ensuring accurate landmark detection under different conditions. Subjectivity vs. objectivity: Beauty is personal, so designing an algorithm that feels meaningful yet respectful was a challenge in itself.
Accomplishments that we're proud of
Built a working AI model that measures beauty ratios in real-time. Successfully integrated computer vision with deep learning for live analysis. Developed a fun, visually interactive app that people can experiment with. Turned a philosophical question—"Can beauty be measured?"—into a functioning software prototype.
What we learned
How to merge mathematical theory (Golden Ratio) with modern AI tools. The importance of balancing performance and accuracy in real-time systems. Insights into the limitations of algorithms when applied to deeply human concepts like beauty. That innovation often comes from experimenting at the intersection of disciplines—art, math, and technology.
What's next for Beauty-ratio(fashionmodels)
Expanding the dataset for more accurate beauty score predictions. Adding gender, age, and cultural calibration for a more inclusive model. Extending applications into fashion, photography, and cosmetic industries for professional use. Creating a mobile app version so anyone can analyze their facial symmetry with just a selfie. Exploring AR integration to give live feedback overlays on facial proportions.
Log in or sign up for Devpost to join the conversation.