StyleM: Stylized Metrics for Image Captioning Built with Contrastive N-grams.

Chengxi Li Brent Harrison

Published in: CoRR (2022)

Keyphrases

n gram
image data
single image
image classification
image representation
language model
bag of words
input image
multiscale
image retrieval
image features
edge detection
image segmentation
variable length
image regions
text classification
databases
viterbi algorithm
language independent
information retrieval