Papers › Self-Supervised Video Representation Learning with Space-Time Cubic Puzzles

Self-Supervised Video Representation Learning with Space-Time Cubic Puzzles

24 Nov 2018arXiv:1811.09795archive 2025-07-28

Dahun Kim, Donghyeon Cho, In So Kweon

Self-supervised tasks such as colorization, inpainting and zigsaw puzzle have been utilized for visual representation learning for still images, when the number of labeled images is limited or absent at all. Recently, this worthwhile stream of study extends to video domain where the cost of human labeling is even more expensive. However, the most of existing methods are still based on 2D CNN architectures that can not directly capture spatio-temporal information for video applications. In this paper, we introduce a new self-supervised task called as \textit{Space-Time Cubic Puzzles} to train 3D CNNs using large scale video dataset. This task requires a network to arrange permuted 3D spatio-temporal crops. By completing \textit{Space-Time Cubic Puzzles}, the network learns both spatial appearance and temporal relation of video frames, which is our final goal. In experiments, we demonstrate that our learned 3D representation is well transferred to action recognition tasks, and outperforms state-of-the-art 2D CNN-based competitors on UCF101 and HMDB51 datasets.

PaperPDF

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

No code repository is listed for this paper in the archive or in Syntology's graph.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Action RecognitionColorizationRepresentation LearningSelf-Supervised Action RecognitionTemporal Action Localization

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Self-Supervised Action Recognition HMDB51 3D Cubic Puzzles (3D ResNet-18) Frozen false #43 of 48 Archive leaderboard report
Self-Supervised Action Recognition HMDB51 3D Cubic Puzzles (3D ResNet-18) Pre-Training Dataset Kinetics400 #43 of 48 Archive leaderboard report
Self-Supervised Action Recognition HMDB51 3D Cubic Puzzles (3D ResNet-18) Top-1 Accuracy 33.7 #43 of 48 Archive leaderboard report
Self-Supervised Action Recognition UCF101 3D Cubic Puzzles (3D ResNet-18) 3-fold Accuracy 65.8 #42 of 53 Archive leaderboard report
Self-Supervised Action Recognition UCF101 3D Cubic Puzzles (3D ResNet-18) Frozen false #42 of 53 Archive leaderboard report
Self-Supervised Action Recognition UCF101 3D Cubic Puzzles (3D ResNet-18) Pre-Training Dataset Kinetics400 #42 of 53 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionAverage PoolingBatch NormalizationBottleneck Residual BlockConvolutionGlobal Average PoolingKaiming InitializationMax PoolingReLUResidual BlockResidual Connection

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections