Upsampling

About: Upsampling is a research topic. Over the lifetime, 2426 publications have been published within this topic receiving 57613 citations.

Scale-Iterative Upscaling Network for Image Deblurring

Abstract: Machine learning based methods for blind deblurring are efficient to handle real-world blurred images, whose blur may be caused by various combined distortions. However, existing multi-level architectures fail to fit images of various scenarios. In this paper, we propose a scale-iterative upscaling network (SIUN) that restores sharp images in an iterative manner. It is not only able to preserve the advantages of weights sharing across scales but also more flexible when training and predicting with different iterations to fit different images. Specifically, we bring in the super-resolution structure instead of the upsampling layer between two consecutive scales to restore a detailed image. Besides, we explore different curriculum learning strategies for both training and prediction of the network and introduce a widely applicable strategy to make SIUN compatible with different scenarios, including text and face. Experimental results on both benchmark datasets and real blurred images show that our method can produce better results than state-of-the-art methods. Code is available at https://github.com/minyuanye/SIUN.

SwinBTS: A Method for 3D Multimodal Brain Tumor Segmentation Using Swin Transformer

Abstract: Brain tumor semantic segmentation is a critical medical image processing work, which aids clinicians in diagnosing patients and determining the extent of lesions. Convolutional neural networks (CNNs) have demonstrated exceptional performance in computer vision tasks in recent years. For 3D medical image tasks, deep convolutional neural networks based on an encoder–decoder structure and skip-connection have been frequently used. However, CNNs have the drawback of being unable to learn global and remote semantic information well. On the other hand, the transformer has recently found success in natural language processing and computer vision as a result of its usage of a self-attention mechanism for global information modeling. For demanding prediction tasks, such as 3D medical picture segmentation, local and global characteristics are critical. We propose SwinBTS, a new 3D medical picture segmentation approach, which combines a transformer, convolutional neural network, and encoder–decoder structure to define the 3D brain tumor semantic segmentation job as a sequence-to-sequence prediction challenge in this research. To extract contextual data, the 3D Swin Transformer is utilized as the network’s encoder and decoder, and convolutional operations are employed for upsampling and downsampling. Finally, we achieve segmentation results using an improved Transformer module that we built for increasing detail feature extraction. Extensive experimental results on the BraTS 2019, BraTS 2020, and BraTS 2021 datasets reveal that SwinBTS outperforms state-of-the-art 3D algorithms for brain tumor segmentation on 3D MRI scanned images.

Improved Point-Voxel Region Convolutional Neural Network: 3D Object Detectors for Autonomous Driving

Abstract: Recently, 3D object detection based on deep learning has achieved impressive performance in complex indoor and outdoor scenes. Among the methods, the two-stage detection method performs the best; however, this method still needs improved accuracy and efficiency, especially for small size objects or autonomous driving scenes. In this paper, we propose an improved 3D object detection method based on a two-stage detector called the Improved Point-Voxel Region Convolutional Neural Network (IPV-RCNN). Our proposed method contains online training for data augmentation, upsampling convolution and k-means clustering for the bounding box to achieve 3D detection tasks from raw point clouds. The evaluation results on the KITTI 3D dataset show that the IPV-RCNN achieved a 96% mAP, which is 3% more accurate than the state-of-the-art detectors.

Three Dimensional Root CT Segmentation Using Multi-Resolution Encoder-Decoder Networks

Abstract: We address the complex problem of reliably segmenting root structure from soil in X-ray Computed Tomography (CT) images. We utilise a deep learning approach, and propose a state-of-the-art multi-resolution architecture based on encoder-decoders. While previous work in encoder-decoders implies the use of multiple resolutions simply by downsampling and upsampling images, we make this process explicit, with branches of the network tasked separately with obtaining local high-resolution segmentation, and wider low-resolution contextual information. The complete network is a memory efficient implementation that is still able to resolve small root detail in large volumetric images. We compare against a number of different encoder-decoder based architectures from the literature, as well as a popular existing image analysis tool designed for root CT segmentation. We show qualitatively and quantitatively that a multi-resolution approach offers substantial accuracy improvements over a both a small receptive field size in a deep network, or a larger receptive field in a shallower network. We then further improve performance using an incremental learning approach, in which failures in the original network are used to generate harder negative training examples. Our proposed method requires no user interaction, is fully automatic, and identifies large and fine root material throughout the whole volume.

Angular spectrum-based wave-propagation method with compact space bandwidth for large propagation distances.

Abstract: Rigorous propagation methods enable diffraction calculations at high NA. However, for the case of large propagation distances and full NA calculations of a signal, common solutions require zero padding or upsampling. This Letter overcomes these problems by introducing a sampling scheme based on compact space bandwidth product representation, which adjusts the sampling frequency of input and propagated field according to the evolution of the generalized space bandwidth product. This sampling concept allows proposing a novel AS method enabling high efficiency, high accuracy, and high-NA diffraction computations at larger propagation distances without need of zero padding or upsampling. The method has several advantages: (1) high accuracy for larger propagation distances; (2) reduced sampling with minimal computation effort; (3) zooming capability; and (4) both focusing and defocusing propagations possible.