Abstract
Deep learning based image compression (DLIC) algorithms have achieved higher compression gain than conventional algorithms. However, the large parameters and float-point operations (FLOPs) of DLIC severely limit their application on mobile devices. To reduce the parameters and FLOPs while maintaining the superior compression gain, this paper proposes lightweight algorithms especially for the feature analysis, synthesis, and fusion modules in DLIC networks. Firstly, based on the observation that there are highly correlated pairing convolution kernels in the analysis/synthesis modules, a new Dynamic Concatenated Convolution (DCC) is proposed to discard half of pairing convolution kernels, which are then restored by the affine transformation of the remaining convolution kernels. Secondly, a novel Depthwise Separable Residual Block (DSRB), utilizing improved depthwise separable convolutions and skipped connections, is proposed to simplify the stacks of residual blocks in feature fusion modules, significantly reducing parameters and FLOPs. Extensive experimental results demonstrate that the proposed lightweight algorithms have fewer parameters/FLOPs and better image compression gain compared with the existing state-of-the-art lightweight algorithms.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Hu, Y., Yang, W., Ma, Z., Liu, J.: Learning end-to-end lossy image compression: a benchmark. IEEE Trans. Pattern Anal. Mach. Intell. 44(8), 4194–4211 (2022)
Ma, Y., Zhai, Y.Q., Yang, J.Y., Yang, C.H., Wang, R.G.: AFEC: adaptive feature extraction modules for learned image compression. In: ACM International Conference on Multimedia, pp. 5436–5444 (2021)
Yang, F., Herranz, L., Weijer, J.V.D., Zhao, J.A.I., López, A.M., Mozerov, M.G.: Variable rate deep image compression with modulated autoencoder. IEEE Signal Process. Lett. 27(2), 331–335 (2020)
Mentzer, F., Agustsson, E., Tschannen, M., Timofte, R., Gool, L.V.: Conditional probability models for deep image compression. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 4394–4402 (2018)
Theis, L., Shi, W., Cunningham, A., Huszár, F.: Lossy image compression with compressive autoencoders (2017). arXiv preprint arXiv:1703.00395
Wallace, G.K.: The jpeg still picture compression standard. Commun. Acm 38(1), xviii-xxxiv (1992)
Christopoulos, C., Skodras, A., Ebrahimi, T.: The JPEG2000 still image coding system: an overview. IEEE Trans. Cons. Electron. 46(4), 1103–1127 (2000)
Bellard, F.: Bpg image format, vol. 1 (2015). https://bellard.org/bpg
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 770–778 (2016)
Cheng, Z., Sun, H., Takeuchi, M., Katto, J.: Deep residual learning for image compression. In: Proceedings of CVPR Workshops (2019)
Yu, R., et al.: Nisp: pruning networks using neuron importance score propagation. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 9194–9203 (2018)
He, Y., Dong, X., Kang, G., Fu, Y., Yan, C., Yang, Y.: Asymptotic soft kernel pruning for deep convolutional neural networks. IEEE Trans. Cybern. 50(8), 3594–3604 (2020)
Tian, G., Chen, J., Zeng, X., Liu, Y.: Pruning by training: a novel deep neural network compression framework for image processing. IEEE Signal Process. Lett. (2021). https://doi.org/10.1109/LSP.2021.3054315
Chollet, F.: Xception: deep learning with depthwise separable convolutions. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 1800–1807 (2017)
Howard, A.G., et al.: Mobilenets: efficient convolutional neural networks for mobile vision applications (2017). arXiv preprint arXiv:1704.04861
Zhang, X., Zhou, X., Lin, M., Sun, J.: Shufflenet: an extremely mefficient convolutional neural network for mobile devices. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 6848–6856 (2018)
Ma, N., Zhang, X., Zheng, H.-T., Sun, J.: Shufflenet v2: practical guidelines for efficient cnn architecture design. In: Proceedings of the European Conference on Computer Vision (ECCV), pp. 116–131 (2018)
Tan, M., Le, Q.V.: Mixconv: mixed depthwise convolutional kernels. arXiv preprint arXiv:1907.09595 (2019)
Daquan, Z., Hou, Q., Chen, Y., Feng, J., Yan, S.: Rethinking bottleneck structure for efficient mobile network design. arXiv preprint arXiv:2007.02269 (2020)
Shang, W., Sohn, K., Almeida, D., Lee, H.: Understanding and improving convolutional neural networks via concatenated rectified linear units. In: Proceedings of International conference on machine learning (ICML), pp. 2217–2225 (2016)
Sandler, M., Howard, A.G., Zhu, M., Zhmoginov, A., Chen, L.C.: MobileNetV2: inverted residuals and linear bottlenecks. In: Proceedings of IEEE/CVF Conference Computer Vision and Pattern Recognition (CVPR), pp. 4510–4520 (2018)
Russakovsky, O., et al.: Imagenet large scale visual recognition challenge. Int. J. Comput. Vision 115(3), 211–252 (2015)
Kodak, E.: The kodak photocd dataset. http://r0k.us/graphics/kodak/.6
Timofte, R., De Smet, V., Van Gool, L.: A+: adjusted anchored neighborhood regression for fast super-resolution. In: Cremers, D., Reid, I., Saito, H., Yang, M.-H. (eds.) ACCV 2014. LNCS, vol. 9006, pp. 111–126. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-16817-3_8
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Li, M., Wang, Z., Shen, L., Ding, Q., Yu, L., Jiang, X. (2022). Lightweight Image Compression Based on Deep Learning. In: Fang, L., Povey, D., Zhai, G., Mei, T., Wang, R. (eds) Artificial Intelligence. CICAI 2022. Lecture Notes in Computer Science(), vol 13604. Springer, Cham. https://doi.org/10.1007/978-3-031-20497-5_9
Download citation
DOI: https://doi.org/10.1007/978-3-031-20497-5_9
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-20496-8
Online ISBN: 978-3-031-20497-5
eBook Packages: Computer ScienceComputer Science (R0)