High-Level Features Parallelization for Inference Cost Reduction Through Selective Attention.
André Peter KelmLucas SchmidtTim RolffChristian WilmsEhsan YaghoubiSimone FrintropPublished in: CoRR (2023)
Keyphrases
- selective attention
- cost reduction
- high level features
- low level features
- focus of attention
- face detection
- video retrieval
- video shots
- selection mechanism
- object recognition
- perceptual grouping
- receptive fields
- bayesian networks
- cost savings
- visual field
- lead time
- visual processing
- high level
- search engine
- multiscale
- natural images
- keywords
- pairwise
- low level
- supply chain