Likelihood Ratio based Loss to finetune CNNs for Very Low Resolution Face Verification.
Dan ZengRaymond N. J. VeldhuisLuuk J. SpreeuwersQijun ZhaoPublished in: ICB (2019)
Keyphrases
- low resolution
- face verification
- likelihood ratio
- face images
- high resolution
- face recognition
- hypothesis testing
- face identification
- super resolution
- face recognition systems
- face detection
- low resolution images
- human faces
- high quality
- lower resolution
- principal component analysis
- high resolution images
- face databases
- image super resolution
- facial features
- facial expressions
- higher resolution
- super resolution reconstruction
- low quality images
- depth images
- recognition rate
- feature extraction
- depth map
- feature vectors
- pose variations
- input image
- training set
- active appearance models
- facial images
- recognition accuracy
- face recognition algorithms
- super resolution algorithm
- computer vision