---
markdown_generated_at: "2026-10-14"
title: "Skip-Convolutions for Efficient Video Processing (CVPR 2021)"
url: "https://www.qualcomm.com/research/artificial-intelligence/papers-with-code/paper-3"
description: "Skip convolutions are a way to save computation and make sure that the neural network focuses only on significant changes in the frame."
canonicalUrl: "https://www.qualcomm.com/research/artificial-intelligence/papers-with-code/paper-3"
---

# Skip-Convolutions for Efficient Video Processing (CVPR 2021)

PAPER 3

Amirhossein Habibian
Davide Abati
Taco S. Cohen
Babak Ehteshami Bejnordi

CVPR21

![](https://s7d1.scene7.com/is/image/dmqualcommprod/papers-3_0?$QC_Responsive$&fmt=png-alpha)

## Summary

Video streams contain many redundancies — in other words, repeated information that is not necessary to process to achieve the same results. Convolutional neural networks process sequences frame by frame, layer by layer. Recalculating this redundant information is extremely compute-inefficient. Skip convolutions are a way to save computation and make sure that the neural network focuses only on significant changes in the frame. For example, if the AI model is focused on tracking the movement of a car, it would skip the frames in which the car stands still.

[![](https://s7d1.scene7.com/is/image/dmqualcommprod/document-icon-v2_2?$QC_Responsive$&fmt=png-alpha) Paper](https://openaccess.thecvf.com/content/CVPR2021/html/Habibian_Skip-Convolutions_for_Efficient_Video_Processing_CVPR_2021_paper.html)

[![](https://s7d1.scene7.com/is/image/dmqualcommprod/code-icon_2?$QC_Responsive$&fmt=png-alpha) Code](https://github.com/Qualcomm-AI-research/Skip-Conv)

![](https://s7d1.scene7.com/is/image/dmqualcommprod/skip-convolutions_for_efficient_video_processing?$QC_Responsive$&fmt=png-alpha)

## Citation

*@inproceedings{skipconv, title={Skip-Convolutions for Efficient Video Processing}, author={Habibian, Amirhossein and Abati, Davide and Cohen, Taco and Bejnordi, Babak Ehteshami}, booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition}, year={2021}}*

## Results

Skip convolutions achieve a significant reduction in computation of 300% to 400%.

[Watch TWiML Interview with Amir Habibian about Skip-Convolutions](https://www.youtube.com/watch?v=ChpuYnWZBss)

## Looking for more papers with code?

[View all papers](https://www.qualcomm.com/research/artificial-intelligence/papers-with-code)

\* Qualcomm AI Research is an initiative of Qualcomm Technologies, Inc.