A novel automatic classification approach for micro-flaws on the large-aperture optics surface based on multi-light source fusion and integrated deep learning architecture.
Zhaoyang YinMingjun ChenLinjie ZhaoJian ChengChao TanXiaotao LiZhiyu LinYankang ChenPublished in: J. Intell. Manuf. (2024)
Keyphrases
- automatic classification
- light source
- deep learning
- object surface
- reflectance properties
- surface orientation
- surface reflectance
- image irradiance
- surface normals
- photometric stereo
- shape from shading
- lambertian surface
- imaging systems
- single image
- lighting conditions
- multiple images
- unsupervised learning
- viewing direction
- automatic detection
- surface reconstruction
- specular reflection
- point light source
- smooth surfaces
- machine learning
- structured light
- mental models
- weakly supervised
- surface shape
- three dimensional
- range data
- occluding contours
- image processing
- human faces
- generative model