Advances in deep learning for real-time image and video reconstruction and processing
- PDF / 334,953 Bytes
- 2 Pages / 595.276 x 790.866 pts Page_size
- 9 Downloads / 222 Views
GUEST EDITORIAL
Advances in deep learning for real‑time image and video reconstruction and processing Pourya Shamsolmoali1 · M. Emre Celebi2 · Ruili Wang3
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Accurate reconstruction algorithms play a vital role in modern imaging techniques. Reconstructing image is a central problem in many key applications including super-resolution imaging, X-ray tomography, ultrasound imaging, remote sensing, and magnetic resonance imaging. The process of image reconstruction typically requires solving an inverse problem that is ill‐posed and large‐scale and thus challenging to solve. The main challenge of this method is its sensitivity to measurement noise in the input data, which will result in artifacts in the reconstructed image with a higher cost in computational time. Thus, it is very important to develop a robust method that can improve reconstruction accuracy while maintaining real-time operation. Real or near real-time processing capabilities are important in image reconstruction techniques for real-world applications. The research field of real-time image reconstruction is very active in image processing and computer vision since it proposes the ability to overcome some of the inherent resolution limitations of low-cost imaging sensors and generates better applications for the emergent capability of highresolution displays. Deep learning for image reconstruction and processing is a relatively new area. Image reconstruction based deep learning can be efficiently performed by using neural networks, in which, weights are determined based on training data. This special issue provides 20 papers * Pourya Shamsolmoali [email protected] M. Emre Celebi [email protected] Ruili Wang [email protected] 1
Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University, Shanghai, China
2
Department of Computer Science, University of Central Arkansas, Conway, AR, USA
3
School of Natural and Mathematical Sciences, Massey University, Auckland, New Zealand
reporting the recent developments of deep learning in image reconstruction. In “Deep Learning Methods in Real-time Image Superresolution: A Survey”, the authors provide a comprehensive survey on real-time image super-resolution. The paper “Investigating Low-Delay Deep Learning-Based Cultural Image Reconstruction”, aims at completing missing regions in artwork through advanced deep learning and image reconstruction (inpainting) techniques. The authors in “A Deep Attention-based Ensemble Network for Real-time Face Hallucination”, propose an end-to-end deep ensemble network that aggregates three sub-networks for extracting attention maps. The paper “A Novel Real-Time Fall Detection Method Based on Head Segmentation and Convolutional Neural Network”, aims at developing a model for fall detection. A gaussian mixture model is proposed for geometric feature extraction to detect the human target and determine the minimum external elliptical contour. In “Effective and Efficient Multitask Learnin
Data Loading...
