Deep Learning: Match vs. Mismatch
Interactive Convolution & Feature Map Generation
1. Patient Image (Input)
Vertical Blood Vessel (Left)
Horizontal Tissue (Middle)
Cross Junction (+)
2. AI Stencil (Filter)
Vertical Edge Detector
Horizontal Edge Detector
Raw Image (5x5)
➜
The Stencil (3x3)
➜
Feature Map (Output)
Select an image and a stencil to see how the AI extracts features.