#ProjectDeepFill
summary
A neural in-painting demo powered by Adobe Sensei that learns to generate image patches from millions of real-world photos and allows user-guided sketch inputs to steer results. It hallucinates missing facial features (recreates eyebrows, noses), removes people from landscapes while keeping semantic coherence, handles multiple masked regions in one pass, and returns several candidate fills for selection. Early research prototype; implementation details point to deep convolutional models trained on large image sets.
AI-written summary · about
chapters (6)
- 0:00Introduction to Project Hill
- 0:18Limitations of current tools
- 1:33AI-powered facial filling
- 2:35Landscape and object removal
- 3:46User-guided sketch filling
- 4:43Technology and conclusion
related talks
56mDeep Fakes 2.0 - How Neural Networks are Changing our World
01h 03mIt's All About AI
58mPushing Deepfakes to the Limit-Fake video calls with AI
56mPicasso's terminal; data science and AI in the visual arts
15mDeepSketch: A New Machine Learning-Based Reference Search Technique...
18mDeep learning for Space imagery applications
25mDecoding Algorithmic Artistry: Visual Generative AI
38mDIY Deep Fakes