AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Dual Directed Capsule Network for Very Low Resolution Image Recognition

Maneet Singh, Shruti Nagpal, Richa Singh, Mayank Vatsa

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Margin-based face recognition losses · 69 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add HR-anchor loss (ablation)DirectCapsNet · SVHN VLR digits (8x8)top-1 accuracy: 82.42 -> 84.51Improved
add HR-anchor loss (ablation)DirectCapsNet · SVHN VLR digits (8x8)top-5 accuracy: 90.15 -> 91.20Improved
add targeted reconstruction loss (ablation)DirectCapsNet · SVHN VLR digits (8x8)top-1 accuracy: 81.95 -> 84.51Improved
add targeted reconstruction loss (ablation)DirectCapsNet · SVHN VLR digits (8x8)top-5 accuracy: 90.35 -> 91.20Improved
add both directed losses to CapsNet trained on VLR imagesCapsNet (VLR) · SVHN VLR digits (8x8)top-1 accuracy: 79.19 -> 84.51Improved
replace CNN with capsule network on VLR imagesCNN (VLR) · SVHN VLR digits (8x8)top-1 accuracy: 45.29 -> 79.19Improved
train CapsNet on VLR images instead of HR images onlyCapsNet · SVHN VLR digits (8x8)top-1 accuracy: 77.82 -> 79.19Improved
use DirectCapsNet instead of SHSRVLR face recognition · SVHN VLR digits (8x8)rank-1 accuracy: 82.6 -> 94.5Improved
use DirectCapsNet instead of Centerloss for VLRUCCS · SVHN VLR digits (8x8)rank-1 accuracy: 93.40 -> 95.81Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy