# Two-Stream Action Recognition-Oriented Video Super-Resolution

arXiv: https://arxiv.org/abs/1903.05577 · HTML: https://mlsanity.com/paper/1903.05577/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Mar 2019.

## Built on by (most cited first, 11 of 11)

- [Learning Deep Context-Sensitive Decomposition for Low-Light Image Enhancement](https://mlsanity.com/paper/2112.05147/) (2021)
- [Reconstructed Convolution Module Based Look-Up Tables for Efficient Image Super-Resolution](https://mlsanity.com/paper/2307.08544/) (2023)
- [Structured Sparsity Learning for Efficient Video Super-Resolution](https://mlsanity.com/paper/2206.07687/) (2022)
- [A Methodological and Structural Review of Hand Gesture Recognition Across Diverse Data Modalities](https://mlsanity.com/paper/2408.05436/) (2024)
- [Siamese Attribute-missing Graph Auto-encoder](https://mlsanity.com/paper/2112.04842/) (2021)
- [Dairy Cow rumination detection: A deep learning approach](https://mlsanity.com/paper/2101.10445/) (2021)
- [Optical Flow for Video Super-Resolution: A Survey](https://mlsanity.com/paper/2203.10462/) (2022)
- [Synergy Between Semantic Segmentation and Image Denoising via Alternate Boosting](https://mlsanity.com/paper/2102.12095/) (2021)
- [Towards Robust and Generalizable Continuous Space-Time Video Super-Resolution with Events](https://mlsanity.com/paper/2510.03833/) (2025)
- [LoLiSRFlow: Joint Single Image Low-light Enhancement and Super-resolution via Cross-scale Transformer-based Conditional Flow](https://mlsanity.com/paper/2402.18871/) (2024)
- [Synchronized and Fine-Grained Head for Skeleton-Based Ambiguous Action Recognition](https://mlsanity.com/paper/2412.14833/) (2024)

## Abstract (arXiv)

We study the video super-resolution (SR) problem for facilitating video analytics tasks, e.g. action recognition, instead of for visual quality. The popular action recognition methods based on convolutional networks, exemplified by two-stream networks, are not directly applicable on video of low spatial resolution. This can be remedied by performing video SR prior to recognition, which motivates us to improve the SR procedure for recognition accuracy. Tailored for two-stream action recognition networks, we propose two video SR methods for the spatial and temporal streams respectively. On the one hand, we observe that regions with action are more important to recognition, and we propose an optical-flow guided weighted mean-squared-error loss for our spatial-oriented SR (SoSR) network to emphasize the reconstruction of moving objects. On the other hand, we observe that existing video SR methods incur temporal discontinuity between frames, which also worsens the recognition accuracy, and we propose a siamese network for our temporal-oriented SR (ToSR) training that emphasizes the temporal continuity between consecutive frames. We perform experiments using two state-of-the-art action recognition networks and two well-known datasets--UCF101 and HMDB51. Results demonstrate the effectiveness of our proposed SoSR and ToSR in improving recognition accuracy.