cs.GRApr 13, 2026

NeuVolEx: Implicit Neural Features for Volume Exploration

Authors: Haill An, Suhyeon Kim, Donghyuk Choo, Younhyun Jung

Organizations: School of Computing, Gachon University, Republic of Korea.

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Direct volume rendering (DVR) aims to help users identify and examine regions of interest (ROIs) within volumetric data, and feature representations that support effective ROI classification and clustering play a fundamental role in volume exploration. Existing approaches typically rely on either explicit local feature representations or implicit convolutional feature representations learned from raw volumes. However, explicit local feature representations are limited in capturing broader geometric patterns and spatial correlations, while implicit convolutional feature representations do not necessarily ensure robust performance in practice, where user supervision is typically limited. Meanwhile, implicit neural representations (INRs) have recently shown strong promise in DVR for volume compression, owing to their ability to compactly parameterize continuous volumetric fields. In this work, we propose NeuVolEx, a neural volume exploration approach that extends the role of INRs beyond volume compression. Unlike prior compression methods that focus on INR outputs, NeuVolEx leverages feature representations learned during INR training as a robust basis for volume exploration. To better adapt these feature representations to exploration tasks, we augment a base INR with a structural encoder and a multi-task learning scheme that improve spatial coherence for ROI characterization. We validate NeuVolEx on two fundamental volume exploration tasks: image-based transfer function (TF) design and viewpoint recommendation. NeuVolEx enables accurate ROI classification under sparse user supervision for image-based TF design and supports unsupervised clustering to identify compact complementary viewpoints that reveal different ROI clusters. Experiments on diverse volume datasets with varying modalities and ROI complexities demonstrate NeuVolEx improves both effectiveness and usability over prior methods

Figures & tables

Figure 1: Overview of our NeuVolEx approach.
Figure 2: An architectural overview of our volume exploration-optimized INR.
Figure 3: User interface of NeuVolEx for image-based TF design.
Figure 4: Visualization results using two variants of the multi-class probabilistic volume renderer: the probabilistic renderer and the probability-intensity renderer.
 F1-score ↑\uparrow
MethodCT-Abdomen (5)CT-Engine (7)CT-Tooth (3)MRA-Brain (1)
Local-Feature0.8001 ±\pm 0.07800.7876 ±\pm 0.07580.8388 ±\pm 0.15110.4335
Convolutional-Feature0.6800 ±\pm 0.15600.7621 ±\pm 0.12750.8271 ±\pm 0.16990.2792
NeuVolEx0.8302 ±\pm 0.06550.8368 ±\pm 0.09870.8795 ±\pm 0.09500.6034
Table 1: Quantitative comparison of ROI classification performance for NeuVolEx against the explicit local-feature method [ 20 ] and the implicit convolutional-feature method [ 19 ] . The number in parentheses denotes the number of ROIs. For datasets with multiple ROI classes, the results indicate the mean and standard deviation over the ROI classes, whereas for MRA-Brain, only the mean is reported because it contains only one ROI class.
 F1-score ↑\uparrow
MethodS1 (756)S2 (1,912)S3 (9,434)S4 (29,529)
Local-Feature0.8247 ±\pm 0.11660.8388 ±\pm 0.15110.8776 ±\pm 0.11290.8900 ±\pm 0.1065
Convolutional-Feature0.7899 ±\pm 0.19270.8271 ±\pm 0.16990.8396 ±\pm 0.15650.8641 ±\pm 0.1287
NeuVolEx0.8648 ±\pm 0.10530.8795 ±\pm 0.09500.8911 ±\pm 0.09670.8946 ±\pm 0.1105
Table 2: Quantitative comparison of scribble sensitivity for NeuVolEx against the explicit local-feature method [ 20 ] and the implicit convolutional-feature method [ 19 ] on CT-Tooth across four scribble levels (S1 to S4). The number in parentheses denotes the number of scribble points; the volume contains 1,558,802 voxels.
Figure 5: Qualitative visualization comparison on four volumes among NeuVolEx, the intensity-TF [ 7 ] , the explicit local-feature method [ 20 ] , and the implicit convolutional-feature method [ 19 ] , using sparse user scribbles.
Figure 6: Qualitative visualization comparison of scribble sensitivity for NeuVolEx, the explicit local-feature method [ 20 ] , and the implicit convolutional-feature method [ 19 ] under four varying scribble levels (S1 – S4) on CT-Tooth.
Figure 7: Qualitative ablation study of NeuVolEx on CT-Abdomen.
Figure 8: Viewpoint recommendation results from NeuVolEx with three volumes
  Feature TrainingClassifier Training
DatasetVolume ResolutionMemory Usage (GB)Processing Time (s)Number of Scribbles (pts)Number of ROIsProcessing Time (s)
CT-Tooth103×94×161103\times 94\times 1611.8511.82191231.75
CT-Bonsai256×256×77256\times 256\times 772.0736.93637342.96
CT-Engine256×256×128256\times 256\times 1282.5061.26597575.96
MRA-Brain332×432×100332\times 432\times 1002.97106.301425712.46
CTA-Head256×256×230256\times 256\times 2303.04112.15836345.47
Table 3: Computational performance of NeuVolEx on the ten evaluation volume datasets with varying resolutions and experiment settings. The bottom three volumes were used for viewpoint recommendation and therefore have no classifier training results.
More in this topic

Implicit Neural Representations

Implicit neural representations encode signals and fields as continuous functions that can be queried at arbitrary coordinates and resolutions.

See all 186 papers

Explore similar work

CardsList
Jul 20, 2026

Lossless-INR: Lossless Volumetric Implicit Neural Representations

Implicit neural representation (INR) methods provide continuous coordinate-to-value mappings and integrate naturally with direct volume rendering, making them attractive for representing volumetric data. However, existing INR-based approaches for volumetric data are inherently lossy, and even small reconstruction errors can propagate through rendering and downstream analysis. In this work, we explore Lossless-INR, a lossless INR framework for 3D scientific volumetric data based on bit-plane decomposition. By decomposing each voxel value into binary bit-planes, we reformulate reconstruction as per-bit binary classification, so that exact recovery reduces to predicting every bit correctly. To make this optimization tractable while keeping the representation compact, we combine an octree block-partitioning strategy that adaptively subdivides complex regions with a ternary feature-grid network whose grid entries are parameterized by a ternary set of values. Experiments on diverse volumetric datasets show that this design can achieve zero bit-error rate and bit-exact reconstruction, enabling faithful rendering and downstream analysis with a compact representation. The code is available at https://github.com/TouKaienn/Lossless-INR.

Kaiyuan Tang, Daniel Burke, Chaoli Wang

University of Notre Dame

Implicit Neural Representations
AskPDF
Jul 30, 2026

A Query-Efficient Stochastic Volume Rendering Framework for Time-Varying Implicit Neural Volumes

Time-varying implicit neural representations (INRs) provide a compact representation of scientific volumes and, for modalities such as dynamic X-ray computed tomography (CT), are often the only practical way to represent the data. However, interactive volume rendering of INRs is challenging, as cheap memory lookups are replaced by expensive neural inferences, hindering the performance. Therefore, conventional volume rendering methods such as ray marching with dense sampling are often impractical. While resampling, caching, and retraining can mitigate this cost, they compromise convenience and accuracy and become impractical for time-varying data. We tackle these challenges using a query-efficient stochastic volume rendering framework based on delta tracking. Our system employs a four-stage pipeline that exploits heterogeneous parallelism, using ray tracing cores for traversal and tensor cores for batched neural evaluation. Furthermore, we present strategies to reduce INR queries via ray budgeting and query pruning, thereby increasing per-frame performance. Using our renderer, many time-varying INRs can be rendered directly from their original representation. The system achieves ~30-40 FPS at 1024x1024 resolution on an RTX 4090 GPU and converges to high-fidelity images. Moreover, the system enables interactive temporal exploration of the continuous domain, with timestep updates taking approximately 1-2 ms.

Alper Sahistan, Haichao Miao, Zhimin Li +3

University of Arizona

Implicit Neural RepresentationsNeural Rendering
AskPDF
Jul 20, 2026

ECoNGS: Efficient Compressive Neural Gaussian Splats for Volume Visualization

Recent advances in differentiable Gaussian splatting have highlighted the potential of primitive-based approaches as alternative scene representations for interactive, high-quality, volume visualization (VolVis) of large datasets. However, the explicit nature of current primitive-based methods, combined with isolated optimization for each VolVis scene, results in redundant, non-compact representations. We present ECoNGS, an efficient compressive neural Gaussian splatting framework for VolVis scene representation. ECoNGS employs lightweight neural networks to dynamically predict implicit, editable Gaussian splats from explicit anchor points, effectively combining model compactness and parameter efficiency of implicit representations with high-performance rendering of explicit primitives. We explore a joint learning strategy that clusters geometrically similar scenes and shares parameters across them, significantly reducing overall training time and model size while maintaining reconstruction fidelity. To achieve a more compact scene representation, we further compress the explicit anchor attributes using a neural entropy model that estimates their probability distributions, enabling compact storage via entropy coding. We systematically investigate Gaussian initialization strategies and propose a simple yet effective scheme tailored for VolVis scenes, improving reconstruction accuracy and accelerating convergence. We evaluate ECoNGS qualitatively and quantitatively across various univariate and multivariate VolVis scenes, highlighting its superior performance over prior methods in training time, reconstruction quality, and model size. In particular, compared with the prior method iVR-GS, ECoNGS improves reconstruction quality by up to 2.2 dB in PSNR while reducing the model size by up to 6.1x and the training time by up to 5.9x. The code is available at https://github.com/TouKaienn/ECoNGS.

Kaiyuan Tang, Chaoli Wang

Department of Computer Science and Engineering, University of Notre Dame, Notre Dame, IN 46556, USA

3DGS CompressionScientific Visualization3D Scene RepresentationNeural Network CompressionGaussian Splatting Compression
AskPDF