LSQ++: Lower Running Time and Higher Recall in Multi-codebook Quantization.
Julieta MartinezShobhit ZakhmiHolger H. HoosJames J. LittlePublished in: ECCV (16) (2018)
Keyphrases
- vector quantization
- significantly lower
- high precision
- significantly higher
- training set
- lower complexity
- data sets
- computational complexity
- image representation
- bag of words
- maximum entropy models
- codebook design
- quantization error
- higher quality
- image compression
- motion estimation
- feature vectors
- case study
- information retrieval