Case Study: VSCO helps users discover the right presets with TensorFlow Lite and ML Kit

VSCO - Customer Case Study

VSCO, a company that provides creative tools and a community for photographers, faced a challenge with user engagement. Their app offered over 160 image-editing presets, but research showed users were overwhelmed by the choice and tended to stick to a few favorites, limiting their creative discovery.

To solve this, VSCO used TensorFlow to build an on-device machine learning feature called "For This Photo." They built a custom deep convolutional neural network model with TensorFlow, converted it for mobile use with TensorFlow Lite, and integrated it into their app using ML Kit. This solution categorizes a user's image in real-time and instantly suggests a curated list of six relevant presets. The feature, which works offline and respects user privacy, was a major success, becoming the second most used preset category in the app.

VSCO