Joint end-to-end CNN training of FFSR super-resolution and RIFE dual-attention extractor yields resolution-invariant ReID features with reported Rank-1 gains of 2.9% and 2.6% on CAVIAR and MLR-CUHK03.
Learning a deep convolutional net- work for image super-resolution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Resolution-invariant Person Re-Identification
Joint end-to-end CNN training of FFSR super-resolution and RIFE dual-attention extractor yields resolution-invariant ReID features with reported Rank-1 gains of 2.9% and 2.6% on CAVIAR and MLR-CUHK03.