The growth stage significantly influences both the yield and quality of Panax notoginseng. Accurate plant phenotypic parameters are crucial for the precise management of P. notoginseng cultivation. Currently, traditio...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
The growth stage significantly influences both the yield and quality of Panax notoginseng. Accurate plant phenotypic parameters are crucial for the precise management of P. notoginseng cultivation. Currently, traditional methods like manual measurements and tools are used for collecting phenotypic information, but they are inefficient and costly. Manual measurements are prone to subjective biases and can potentially harm plants irreversibly. While 2D images can capture plant phenotypic information, they often suffer fromincompletedata and lack precision, especially for plants with complex structures like P. notoginseng. This limitation makes it challenging to achieve comprehensive and accurate measurements. To address these challenges, this study proposes a novel approach using a Neural Radiance Field (NeRF) for extracting phenotypic parameters from P. notoginseng. By capturing video and multi-view image sequences of P. notoginseng, we were able to achieve high-fidelity 3D rendering of the plants and extract point cloud datafrom them. This approach enabled accurate measurement of plant height and leaf area parameters. The results demonstrate promising accuracy, with an average percentage error of 1.76% for plant height and 1.73% for leaf area based on the point cloud measurements obtained using NeRF. This method leverages advanced computational techniques to overcome the limitations of traditional 2D imaging methods, offering a more comprehensive and precise means of phenotypic characterization for complex plant structures like P. notoginseng.
Multi-spectral optoacoustic tomography (MSOT) relies on optical excitation of tissues with subsequent detection of the generated ultrasound waves. Optimal image quality in MSOT is achieved by detection of signals from...
详细信息
Multi-spectral optoacoustic tomography (MSOT) relies on optical excitation of tissues with subsequent detection of the generated ultrasound waves. Optimal image quality in MSOT is achieved by detection of signals from a broad tomographic view. However, due to physical constraints and other cost-related considerations, most imaging systems are implemented with probes having limited tomographic coverage around the imaged object, such as linear array transducers often employed for clinical ultrasound (US) imaging. MSOT imagereconstructionfrom limited-view data results in arc-shaped image artifacts and disrupted shape of the vascular structures. Deep learning methods have previously been used to recover MSOT images fromincomplete tomographic data, albeit poor performance was attained when training with datafrom simulations or other imaging modalities. We propose a two-step method consisting of i) style transfer for domain adaptation between simulated and experimental MSOT signals, and ii) supervised training on simulated data to recover missing tomographic signals in realistic clinical data. The method is shown capable of correcting images reconstructed from sub-optimal probe geometries using only signal domain data without the need for training with ground truth (GT) full-view images.
Magnetic resonance imaging (MRI) acceleration is usually achieved by data undersampling, while reconstructionfrom undersampled data is a challenging ill-posed problem for data-missing and noisy measurements introduce...
详细信息
ISBN:
(纸本)9783031164460;9783031164453
Magnetic resonance imaging (MRI) acceleration is usually achieved by data undersampling, while reconstructionfrom undersampled data is a challenging ill-posed problem for data-missing and noisy measurements introduce various artifacts. In recent years, deep learning methods have been extensively studied for MRI reconstruction, and most of work treat the reconstruction problem as a denoising problem or replace the regularization subproblem with a deep neural network (DNN) in an optimization unrolling scheme. In this work, we proposed to directly complete the missing and corrupted k-space data by a specially designed interpolation deep neural networks combined with some convolution layers in both frequency and spatial domains. Specifically, for every missing and corrupted frequency, we use a K- nearest neighbors estimation with learnable weights. Then, two convolution neural networks (CNNs) are applied to regularize the data in both k-space and image space. The proposed DNN structures have clear interpretability for solving this undersampling problem. Extensive experiments on MRI reconstruction with diverse sampling patterns and ratios, under noiseless and noise settings demonstrate the accuracy of the proposed method compared to other learning based algorithms, while being computationally more efficient for both training and reconstruction processes.
Reconstructing undersampled medical images from k-space magnetic resonance imaging (MRI) data with multi-coil acquisitions is a major challenge for deep learning algorithms. Applying convolutional architecture has bec...
详细信息
We present a method for the accurate 3D reconstruction of partly-symmetric objects. We build on the strengths of recent advances in neural reconstruction and rendering such as Neural Radiance Fields (NeRF). A major sh...
详细信息
ISBN:
(纸本)9783031198236;9783031198243
We present a method for the accurate 3D reconstruction of partly-symmetric objects. We build on the strengths of recent advances in neural reconstruction and rendering such as Neural Radiance Fields (NeRF). A major shortcoming of such approaches is that they fail to reconstruct any part of the object which is not clearly visible in the training image, which is often the case for in-the-wild images and videos. When evidence is lacking, structural priors such as symmetry can be used to complete the missing information. However, exploiting such priors in neural rendering is highly non-trivial: while geometry and non-reflective materials may be symmetric, shadows and reflections from the ambient scene are not symmetric in general. To address this, we apply a soft symmetry constraint to the 3D geometry and material properties, having factored appearance into lighting, albedo colour and reflectivity. We evaluate our method on the recently introduced CO3D dataset, focusing on the car category due to the challenge of reconstructing highly-reflective materials. We show that it can reconstruct unobserved regions with high fidelity and render high-quality novel view images.
Augmented Reality (AR) applications demand realistic rendering of virtual content in a variety of environments, so they require an accurate description of the 3-D scene. In most case AR system is equipped with Time-of...
详细信息
ISBN:
(纸本)9781510655546
Augmented Reality (AR) applications demand realistic rendering of virtual content in a variety of environments, so they require an accurate description of the 3-D scene. In most case AR system is equipped with Time-of-Flight (ToF) cameras to provide real-time scene depth maps, but they have problems that affect the quality of depth data, which ultimately makes them difficult to use for AR. Such defects appear because of poor lighting, specular or fine-grained surfaces of objects. As a result, the effect of increasing the boundaries of objects appears, and the overlapping of objects makes it impossible to distinguish one object from another. The article presents an approach based on a modified algorithm for searching for similar blocks using the concept of anisotropic gradient. A proposed modified exemplar block-based algorithm uses the autoencoder-learned local image descriptor for image inpainting, that extract the features of images, and the depth image by a decoding network. The encoder consists of a convolutional layer and a dense block, which also consists of convolutional layers. We also show the application for the proposed vision system using depth inpainting for virtual content reconstruction in augmented reality. Analysis of the results of the study shows that the proposed method allows you to correctly restore the boundaries of objects on the image of the depth map. Our system quantitatively outperforms state-of-the-art methods in terms of reconstruction accuracy in the real and simulated benchmark datasets.
Unrolled neural networks have enabled state-of-the-art reconstruction performance and fast inference times for the accelerated magnetic resonance imaging (MRI) reconstruction task. However, these approaches depend on ...
详细信息
ISBN:
(纸本)9783031164460;9783031164453
Unrolled neural networks have enabled state-of-the-art reconstruction performance and fast inference times for the accelerated magnetic resonance imaging (MRI) reconstruction task. However, these approaches depend on fully-sampled scans as ground truth data which is either costly or not possible to acquire in many clinical medical imaging applications;hence, reducing dependence on data is desirable. In this work, we propose modeling the proximal operators of unrolled neural networks with scale-equivariant convolutional neural networks in order to improve the data-efficiency and robustness to drifts in scale of the images that might stem from the variability of patient anatomies or change in field-of-view across different MRI scanners. Our approach demonstrates strong improvements over the state-of-the-art unrolled neural networks under the same memory constraints both with and without data augmentations on both in-distribution and out-of-distribution scaled images without significantly increasing the train or inference time.
Compressive imaging (CI) consists of reconstructing images fromincomplete observed data. The reconstruction process involves solving an ill-posed inverse problem which is highly dependent on the number of real measur...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
Compressive imaging (CI) consists of reconstructing images fromincomplete observed data. The reconstruction process involves solving an ill-posed inverse problem which is highly dependent on the number of real measurements, with a greater number of measurements typically leading to more accurate reconstructions. Due to their ability to learn data distributions, diffusion models (DM) have emerged as promising techniques for various inverse problems. Mainly, DMs solve inverse problems by conditioning the generation process to the acquired measurements. In this work, we introduce a new approach to improve this conditioning by exploiting synthetic measurements, which come from a synthetic sensing matrix. Synthetic measurements are estimated from real datavia a neural network. The combined real and synthetic measurements form an augmented set, which is input into the conditional DM to enhance reconstruction capacity. Computational experiments demonstrate that augmenting measurements with the conditional DM improves performance compared to using only real measurements.
Working with observational data in the context of geophysics can be challenging, since we often have to deal with missing data. This requires imputation techniques in pre-processing to obtain data-mining-ready samples...
详细信息
The proceedings contain 40 papers. The special focus in this conference is on German Association for Pattern Recognition. The topics include: Airborne-Shadow: Towards Fine-Grained Shadow Detection in Aerial image...
ISBN:
(纸本)9783031546044
The proceedings contain 40 papers. The special focus in this conference is on German Association for Pattern Recognition. The topics include: Airborne-Shadow: Towards Fine-Grained Shadow Detection in Aerial imagery;UGainS: Uncertainty Guided Anomaly Instance Segmentation;local Spherical Harmonics Improve Skeleton-Based Hand Action Recognition;LMD: Light-Weight Prediction Quality Estimation for Object Detection in Lidar Point Clouds;a Network Analysis for Correspondence Learning via Linearly-Embedded Functions;HiFiHR: Enhancing 3D Hand reconstructionfrom a Single imagevia High-Fidelity Texture;point2Vec for Self-supervised Representation Learning on Point Clouds;fullFormer: Generating Shapes Inside Shapes;GenLayNeRF: Generalizable Layered Representations with 3D Model Alignment for Human view Synthesis;RC-BEVFusion: A Plug-In Module for Radar-Camera Bird’s Eye view Feature Fusion;parallax-Aware image Stitching Based on Homographic Decomposition;dustNet: Attention to Dust;leveraging Bioclimatic Context for Supervised and Self-supervised Land Cover Classification;Automatic Reverse Engineering: Creating Computer-Aided Design (CAD) Models from Multi-view images;characterization of Out-of-distribution Samples from Uncertainty Maps Using Supervised Machine Learning;underwater Multiview Stereo Using Axial Camera Models;3D Retinal Vessel Segmentation in OCTA Volumes: Annotated dataset MORE3D and Hybrid U-Net with Flattening Transformation;m(otion)-Mode Based Prediction of Ejection Fraction Using Echocardiograms;improving data Efficiency for Plant Cover Prediction with Label Interpolation and Monte-Carlo Cropping;learning Channel Importance for High Content Imaging with Interpretable Deep Input Channel Mixing;self-supervised Learning in Histopathology: New Perspectives for Prostate Cancer Grading;deviL: Decoding vision features into Language;Zero-Shot Translation of Attention Patterns in VQA Models to Natural Language;beyond Debiasing: Actively Steering Feature Selection vi
暂无评论