AI Paper Atlas

Research map / Video / Action recognition

RubiksNet: Learnable 3D-Shift for Efficient Video Action Recognition

Linxi Fan, Shyamal Buch, Guanzhi Wang, Ryan Cao, Yuke Zhu, Juan Carlos Niebles, Li Fei-Fei

ECCV 2020 · Video › Action recognition · approach: Spatiotemporal CNNs for action recognition · 65 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
learn spatial shifts instead of fixed [29] (ablation)RubiksNet-Large · HMDB-51top-1 accuracy: 70.0 -> 71.4Improved
learn temporal shifts instead of fixed [16] (ablation)RubiksNet-Large · HMDB-51top-1 accuracy: 69.5 -> 71.4Improved
learn both spatial and temporal shifts instead of both fixed (ablation)RubiksNet-Large · HMDB-51top-1 accuracy: 68.3 -> 71.4Improved
learn both spatial and temporal shifts instead of both fixed (ablation)RubiksNet-Small · HMDB-51top-1 accuracy: 65.7 -> 69.5Improved
learn temporal shifts instead of fixed [16] (ablation)RubiksNet-Small · HMDB-51top-1 accuracy: 67.8 -> 69.5Improved
learn spatial shifts instead of fixed [29] (ablation)RubiksNet-Small · HMDB-51top-1 accuracy: 67.5 -> 69.5Improved
use interpolated quantized (RS-IQ) instead of interpolated shift (ablation)RubiksShift · Val Top-1 (shift type ablation)top-1 accuracy: 61.7 -> 58.2Hurt
use attention quantized (RS-AQ) with 0.125 budget instead of full interpolated (ablation)RubiksShift · Val Top-1 (shift type ablation)top-1 accuracy: 61.7 -> 61.6No clear effect
use attention quantized (RS-AQ) instead of interpolated quantized (RS-IQ) (ablation)RubiksShift · Val Top-1 (shift type ablation)top-1 accuracy: 58.2 -> 61.6Improved
use RubiksNet-Tiny instead of TSM-Smallvideo recognition · Something-Something-v2 (1-clip)top-1 accuracy: 49.3 -> 54.6Improved
use RubiksNet-Large instead of TSM-Largevideo recognition · Something-Something-v2 (1-clip)top-1 accuracy: 58.8 -> 59.0No clear effect
use RubiksNet instead of TSMvideo recognition · HMDB-51 (2-clip, Large)top-1 accuracy: 73.5 -> 74.6Improved

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