What is Local Binary Pattern (LBP), and why is it well-suited to face recognition?
LBP is a texture descriptor that splits a face into regions, extracts a characteristic texture pattern from each, and combines them into one global descriptor — robust to lighting changes and computationally cheap.
Local Binary Pattern (LBP) is one of the most powerful texture descriptors in image processing. For faces it works in three steps:
- Regional segmentation: the face image is divided into local regions, each analysed independently.
- Texture extraction: a specific texture description capturing characteristic patterns is extracted from each region.
- Global description: the regional descriptions are concatenated into one global face descriptor.
Technical advantage: LBP is robust to illumination changes and computationally efficient, which makes real-time applications possible.
Tip: Because LBP encodes local contrast patterns rather than absolute brightness, a face stays recognisable whether it's lit from the left or right — that lighting-invariance is its killer feature.
Go deeper:
Local binary patterns (Wikipedia) — the texture-descriptor algorithm and its lighting-robust face-recognition use.