Abstract
Recently, deep learning has become a hot topic in wide fields, especially in the computer vision that proved its efficiency in processing images. However, it tends to overfit or consumes a long learning time in many platforms. The causes behind these issues return to the huge number of learning parameters and lack or incorrect training samples. In this work, two levels of deep convolutional neural network (DCNN) are proposed for classifying the images. The first one is enhancing the training images with removing unnecessary details, and the second one is detecting the edges of the processed images for further reduction of learning time in the DCNN. The proposed work is inspired by the human eye's way in recognizing an object, where a piece of object can be helpful in the recognition and not necessarily the whole object or full colors. The goal is to speed up the learning process of CNN based on the preprocessed training samples that are precise and lighter to work well in real-time applications. The obtained results proved to be more significant for real-time classification as it reduced the learning process by (94%) in Animals10 dataset with a validation accuracy of (99.2%) in accordance with the classical DCNNs.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.References
Schmidhuber, J.: Deep learning in neural networks: an overview. Neural Netw. 61, 85–117 (2015)
Coates, A., et al., An analysis of single-layer networks in unsupervised feature learning. In: Proceedings 14th International Conference on Artificial Intelligence and Statistics (AISTATS). 2011, Fort Lauderdale, FL, USA. Vol. 15 of JMLR: W&CP 15 (2011)
N. B. Aldin, et.al., Accuracy comparison of different batch size for a supervised machine learning task with image classification. In: 2022 9th International Conference on Electrical and Electronics Engineering (ICEEE), pp. 316–319 (2022). https://doi.org/10.1109/ICEEE55327.2022.9772551
Vesal, S., et.al, (2018). Classification of breast cancer histology images using transfer learning. In: ICIAR. Lecture Notes in Computer Science, vol. 10882. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-93000-8_92
Collobert, R., et al.: Natural language processing (almost) from scratch. JMLR 12(1), 2493–2537 (2011)
Tang, W., He, F., Liu, Y., Duan, Y.: MATR: multimodal medical image fusion via multiscale adaptive transformer. IEEE Trans. Image Process. 31, 5134–5149 (2022). https://doi.org/10.1109/TIP.2022.3193288
Tang, W., He F., Liu, Y.: YDTR: infrared and visible image fusion via Y-shape dynamic transformer. In: IEEE Transactions on Multimedia, (2022). https://doi.org/10.1109/TMM.2022.3192661
Zhang, S., He, F.: DRCDN: learning deep residual convolutional dehazing networks. Vis Comput 36, 1797–1808 (2020). https://doi.org/10.1007/s00371-019-01774-8
Si, T., He, F., Zhang, Z., Duan, Y., Hybrid contrastive learning for unsupervised person re-identification. In: IEEE Transactions on Multimedia. https://doi.org/10.1109/TMM.2022.3174414
LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015)
Xie, S., et al.: Aggregated residual transformations for deep neural networks. In: Proceedings of the IEEE conference on CVPR. (2017)
Ioffe, S., et al.: Batch normalization: accelerating deep network training by reducing internal covariate shift, In: Proceedings of the 2015 International Conference on MLIS, Gangzhou, China (2015)
Srivastava, N., et al.: Dropout: a simple way to prevent neural networks from overfitting. JMLR 15(1), 1929–1958 (2014)
Zhang, H., et al.: mixup: beyond empirical risk minimization. (2017). https://arxiv.org/abs/1710.09412
Yun, S., et al., Cutmix: regularization strategy to train strong classifiers with localizable features. (2019). https://arxiv.org/abs/1905.04899
Zhong, Z., et al.: Random erasing data augmentation. (2017). https://arxiv.org/abs/1708.04896
Yang, F., Li, B.: Unsupervised learning of spatial structures shared among images. Vis. Comput. 28(2), 175–180 (2012)
Zhang, L., et al.: Extracting shared subspace incrementally for multilabel image classification. Vis. Comput. 30, 1359–1371 (2014). https://doi.org/10.1007/s00371-013-0891-4
Huang, K., et al.: Image saliency detection via multi-scale iterative CNN. In: The Visual Computer, pp. 1–13 (2019)
Liu, Z., Xiang, Q., Tang, J., et al.: Robust salient object detection for RGB images. Vis. Comput. 36, 1823–1835 (2020). https://doi.org/10.1007/s00371-019-01778-4
Carsten Rother, et al.: Grabcut: Interactive foreground extraction using iterated graph cuts. In: ACM TOG, vol. 23, pp. 309–314. ACM (2004)
Xiaodan Liang, et al.: Reversible recursive instance-level object segmentation. In: The IEEE Conference on CVPR, June (2016)
Macoadha, O., Github, Inception v3 model (2018). https://github.com/macaodha/inat_comp_2018/
Wan, L., et al.: Regularization of neural networks using dropconnect, In: Proceedings of the 30th, ICML, pp. 1058–1066, June (2013)
Krizhevsky, A. et al.: Imagenet classification with deep CNNs, In: Proceedings of the 26th Annual Conference on NeurIPS, pp. 1106–1114, USA, December (2012)
Julie P.: GrabCut for Automatic Image Segmentation [OpenCV Tutorial]. Sicara, 26 November 2020. https://www.sicara.ai/blog/grabcut-for-automatic-image-segmentation-opencv-tutorial
Huang, G., et al.: Densely connected convolutional networks. In: IEEE Conference on CVPR, pp. 2261–2269 (2017).
M. J. Huiskes, et al.: The MIR flickr retrieval evaluation. ACM ICMR, (MIR'08), Vancouver, Canada (2008)
Julie P.: GrabCut for Automatic Image Segmentation [OpenCV Tutorial]. Sicara, 26 November 2020. https://www.sicara.ai/blog/grabcut-for-automatic-image-segmentation-opencv-tutorial
Alessio, C.: Animals-10. Kaggle, 2019, https://www.kaggle.com/datasets/alessiocorrado99/animals10. Accessed 26 August 2020.
Sparsh, G.: Flowers Dataset. Kaggle, 2020 https://www.kaggle.com/datasets/imsparsh/flowers-datase. Accessed 26 August 2020
Roy, P., et al.: Natural images. Kaggle. (2018). https://www.kaggle.com/datasets/prasunroy/natural-images. Accessed 26 Aug 2020
Kane, Aditya. “Imagenette2.”, Kaggle. (2019). https://www.kaggle.com/datasets/adityakane/imagenette2. Accessed 26 Aug 2020.
NEOZ, RO.: Kaggle. (2020) https://www.kaggle.com/datasets/roneoz/imagenette160. Accessed 26 Aug 2020
SACHIN. “Cats-vs-Dogs.” Kaggle. (2020). https://www.kaggle.com/datasets/shaunthesheep/microsoft-catsvsdogs-dataset. Accessed 26 Aug 2020
Kang, G., et al.: Patchshuffle regularization. (2017). https://arxiv.org/abs/1707.07103
Devries, T., et al.: Improved regularization of CNNs with cutout. (2017). https://arxiv.org/abs/1708.04552
Nazeri, K., et al.: Two-stage CNN for breast cancer histology image classification In: Proceedings of the 15th, ICIAR 2018, pp. 717–726. P´ovoa de Varzim, Portugal (2018)
Aresta, G., et al.: BACH: grand challenge on breast cancer histology images. Med. Image Anal. 56, 122–139 (2019)
He, K., et al.: Deep residual learning for image recognition. In: Proceedings of the IEEE conference on CVPR, pp. 770–778 (2016)
Srivastava, et al.: Learning representations for multimodal data with deep belief nets. In: International Conference on Machine Learning Workshop, vol. 79 (2012)
Wang, R., et al.: Large scale automatic image annotation based on convolutional neural network. J. Vis. Comm. Image Rep. 49, 213–224 (2017)
Funding
The authors have no financial or proprietary interests in any material discussed in this article.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors did not receive support from any organization for the submitted work.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Safa aldin, S., Aldin, N.B. & Aykaç, M. Enhanced image classification using edge CNN (E-CNN). Vis Comput 40, 319–332 (2024). https://doi.org/10.1007/s00371-023-02784-3
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00371-023-02784-3