---
title: 'Style4D-Bench: 4D Stylization Benchmark'
url: https://www.emergentmind.com/topics/style4d-bench
type: topic
---

# Style4D-Bench: 4D Stylization Benchmark

Style4D-Bench is a benchmark suite for 4D stylization, defined as stylizing dynamic 3D scenes over time—\(4\mathrm{D} = 3\mathrm{D}\) space \(+ 1\mathrm{D}\) time—so that rendered outputs from arbitrary viewpoints and times preserve spatial fidelity, temporal coherence, and multi-view consistency. It is presented as the first benchmark suite specifically designed for 4D stylization and combines a benchmark protocol, a curated set of high-resolution dynamic scenes, and a baseline method, Style4D, built upon 4D Gaussian Splatting (4DGS) [2508.19243].

## 1. Scope and problem formulation

In this setting, **spatial fidelity** denotes preservation of geometry and structural details of the original scene; **temporal coherence** denotes the avoidance of flickering or jitter across time; and **multi-view consistency** denotes stylistic consistency across viewpoints. The benchmark therefore targets a joint problem: stylized rendering of dynamic 3D scenes under both view variation and temporal evolution [2508.19243].

The paper distinguishes this problem from two adjacent regimes. In **2D video stylization**, computation occurs in image space; temporal constraints may be imposed, but the methods lack explicit 3D geometric reasoning and therefore cannot guarantee multi-view consistency or geometry-faithful stylization for novel views. In **3D/NeRF stylization**, multi-view consistency can be enforced for static scenes, but temporal dynamics are typically ignored, so temporal coherence under motion, deformation, and occlusion is not guaranteed. **4D stylization** is therefore defined by the simultaneous requirement to satisfy 3D multi-view constraints and 1D temporal stability on dynamic scenes.

Within this formulation, 4DGS functions as the geometry backbone. The benchmark describes 4DGS as extending 3D Gaussian Splatting to dynamic scenes by combining canonical geometry and appearance with a deformation field over time. In Style4D, this provides both accurate dynamic-scene reconstruction and a differentiable renderer whose appearance can be modulated for stylization while preserving geometry and enabling consistent multi-view rendering. This suggests that the benchmark treats geometry preservation not as an auxiliary desideratum but as a central precondition for evaluating stylization in dynamic 3D scenes.

## 2. Benchmark composition and data regime

Style4D-Bench is instantiated on the Neu3D dataset, curated here as dynamic 4D scenes for stylization evaluation. The benchmark contains **6 dynamic scenes**: `cook_spinach`, `flame_salmon_1`, `sear_steak`, `flame_steak`, `coffee_martini`, and `cut_roasted_beef`. Each scene contains **15–20 static cameras** distributed in space, **300 frames** per scene, and a frame resolution of **\(1352 \times 1014\)**. The scenes exhibit diverse motions, including cooking actions and pouring, together with complex backgrounds, nontrivial viewpoint variation, and occlusion [2508.19243].

Evaluation uses both **fixed test viewpoints** and **helical/spiral novel-view trajectories** rendered across time. The fixed viewpoints probe held-out-view generalization, while the helical or spiral trajectories probe multi-view consistency and temporal stability under continuous camera motion. The use of long sequences at native resolution is a defining property of the suite.

Style specification is image-based rather than prompt-based. Style images are sampled from **WikiArt** to define target styles, and content images used to train the 2D stylization module are drawn from **COCO2014**. The paper evaluates similarity between stylized outputs and the target style using **CKDN** and **LPIPS-to-style**. It does not enumerate fixed style categories, and **no text prompts are used**.

The project page is listed as `https://becky-catherine.github.io/Style4D/`. The dynamic scene data derive from Neu3D, and usage follows Neu3D’s license and terms. The paper states that the benchmark protocol, baseline code pointers, and evaluation details are provided via the project page.

## 3. Evaluation protocol and metric decomposition

The benchmark decomposes evaluation into **“4D stylization quality”** and **“4D stylization consistency”**, comprising **six dimensions and 12 metrics**. Scores are computed per frame and averaged over sequences; for multi-view assessment, they are averaged over frames along spiral trajectories. The paper explicitly states that it does **not** use **FID** or **CLIPScore**, and that **PSNR** is used only in an auxiliary reconstruction ablation, not as a primary stylization metric [2508.19243].

| Dimension | Metrics | Direction |
|---|---|---|
| Imaging Quality | UIQM, CLIP-IQA+, MUSIQ | Higher is better |
| Aesthetic Quality | Q-Align, MUSIQ-PAQ2PIQ | Higher is better |
| Spatiotemporal Consistency | DISTS, Warp loss | DISTS lower, Warp lower |
| Subject Consistency | Inter-frame DINO feature similarity | Higher is better |
| Style Consistency | CKDN, LPIPS-to-style | CKDN higher, LPIPS lower |
| Content Consistency | SSIM, LPIPS-to-content | SSIM higher, LPIPS lower |

**Imaging quality** is measured with **UIQM**, **CLIP-IQA+**, and **MUSIQ**. UIQM is described as a no-reference score combining colorfulness, sharpness, and contrast. CLIP-IQA+ is a CLIP-based no-reference image quality assessment measure. MUSIQ is identified as a multi-scale transformer-based NIQE. **Aesthetic quality** is measured with **Q-Align**, an LMM-based discrete-level aesthetic assessment, and **MUSIQ-PAQ2PIQ**, trained on PaQ-2-PiQ for aesthetics.

**Spatiotemporal consistency** is assessed with **DISTS** and an optical-flow-based **Temporal Warping Error** on helical-trajectory videos. Using consecutive frames \(I_t\) and \(I_{t+1}\), and RAFT optical flow \(F_{t \to t+1}\), the benchmark defines:
$$
\hat{I}_{t+1}(x) = I_{t+1}(x + F_{t \to t+1}(x))
$$
$$
L_{\mathrm{warp}}^t = \frac{1}{H W}\sum_x \left| \hat{I}_{t+1}(x) - I_t(x) \right|
$$
$$
L_{\mathrm{warp}} = \frac{1}{T-1}\sum_{t=1}^{T-1} L_{\mathrm{warp}}^t
$$
DISTS is used as a reference-based perceptual similarity between consecutive frames, with lower values preferred when comparing frame-to-warped-frame. **Subject consistency** is measured by inter-frame **DINO** feature similarity at a fixed viewpoint.

For **style consistency**, the suite uses **CKDN** and **LPIPS-to-style**. For **content consistency**, it uses **SSIM** and **LPIPS-to-content**, both computed against the original content frame from the same camera and time.

Human perceptual evaluation supplements the metric suite. The study uses **34 participants** in two parts. In the **single-frame** study, participants evaluate **5 pairs of images**, with **four extracted frames per video**, rating stylization quality and image quality. In the **long-video** study, participants evaluate **two 10-second pairs** containing **300 frames each**, rating stylization quality, spatiotemporal consistency, and video quality. Results are reported as voting percentages.

## 4. Style4D baseline architecture

To establish a strong baseline, the paper introduces **Style4D**, a 4D stylization framework built upon 4DGS. The framework contains three components: a **basic 4DGS scene representation** to capture reliable geometry, a **Style Gaussian Representation** using lightweight per-Gaussian MLPs for temporally and spatially aware appearance control, and a **Holistic Geometry-Preserved Style Transfer** module (**HGST**) designed to enhance spatio-temporal consistency via contrastive coherence learning and structural content preservation [2508.19243].

### Base 4DGS representation

The canonical or static Gaussians per scene are defined as
$$
G_i = \{ \mu_i, r_i, s_i, o_i, c_i^{sh} \},
$$
where \(\mu_i \in \mathbb{R}^3\) is the center, \(r_i\) the rotation, \(s_i\) the scale, \(o_i\) the opacity, and \(c_i^{sh}\) the spherical-harmonic coefficients for view-dependent color. A deformation field network \(F_{\mathrm{def}}\) warps the Gaussians across time.

Rendering uses front-to-back alpha compositing of rasterized Gaussian splats:
$$
C(p) = \sum_{i \in N(p)} T_i \alpha_i(p)c_i,
\qquad
T_i = \prod_{j<i}(1-\alpha_j(p)),
$$
where \(N(p)\) is the depth-sorted neighborhood of Gaussians contributing to pixel \(p\).

### Style Gaussian Representation

Each Gaussian receives a lightweight, per-Gaussian MLP with a style code \(f_i^{\mathrm{style}}\) that modulates appearance over space-time while preserving geometry. The inputs include time \(t\) and the ray–ellipsoid intersection depth \(p_t\), computed from pixel \(p\) and camera pose \(M = [R, T]\), following the 3D/4DGS ray–Gaussian intersection. Rendering with style modulation is defined as:
$$
c(p) = \sum_{i \in N(p)} \left(c_i + F_c(p_t,t)\right)\cdot F_\alpha^i \cdot \prod_{j=1}^{i-1}(1-\alpha_j(p)).
$$
Here \(c_i\) is the base view-dependent color, \(F_c\) is the MLP-predicted color increment, and \(F_\alpha^i\) together with \(\alpha_j(p)\) denotes opacities modulated per Gaussian by the MLPs.

The paper states that this formulation preserves 3D geometry because modulation is applied at intersections and uses depth and time, enables temporally aware stylization, and improves multi-view consistency. In the reported experiments, each per-Gaussian MLP consists of **two layers**. The appendix specifies that the MLP takes **time** and **intersection depth**, uses **4 hidden units**, and outputs **4 channels** corresponding to **RGB deltas and opacity**.

### Holistic Geometry-Preserved Style Transfer

HGST is an encoder–transformer–decoder stylization backbone that stylizes 2D frames while preserving structure and coherence. It fuses content and style via multichannel correlation and is regularized by two key consistency terms.

The **local contrastive loss** aligns local differences using an attention-guided design with CBAM attention. For features at scales \(i = 3..5\), after sampling \(N\) locations \(G_x\) and their \(8\)-NN \(G_y\), and defining local differences \(d^g_{x,y} = G_x^{fcs}_i - G_{x,y}^{fcs}_i\) and \(d^c_{x,y} = G_x^{fc}_i - G_{x,y}^{fc}_i\), the loss is:
$$
L_{\mathrm{lcl}} = \sum_{m=1}^{8N}
-\log
\left[
\frac{\exp\left((d_m^g \cdot d_m^c)/\tau\right)}
{\sum_{n=1}^{8N}\exp\left((d_m^g \cdot d_n^c)/\tau\right)}
\right],
\qquad \tau = 0.07.
$$

The **global content loss** preserves structure:
$$
L_{\mathrm{content}} = \frac{1}{N}\sum_{i=1}^{N}\lVert f_i^{cs} - f_i^{c} \rVert_2^2.
$$

These define the combined consistency loss:
$$
L_{\mathrm{consistency}} = L_{\mathrm{lcl}} + L_{\mathrm{content}}.
$$

Additional stylization objectives are also included:

- **Style perceptual loss**:
$$
L_{\mathrm{style}} =
\sum_l
\left[
\lVert \mu_l(I_{cs}) - \mu_l(I_s)\rVert_2^2 +
\lVert \sigma_l(I_{cs}) - \sigma_l(I_s)\rVert_2^2
\right]
$$
- **Identity loss**:
$$
L_{\mathrm{id}} = \lVert I_{cc} - I_c \rVert^2 + \lVert I_{ss} - I_s \rVert^2
$$
- **Illumination loss**:
$$
L_{\mathrm{illum}} = \lVert G(I_c, I_s) - G(I_c + \epsilon, I_s) \rVert^2,
\qquad
\epsilon \sim \mathcal{N}(0,\sigma^2 I)
$$
- **Inner-channel similarity loss** \(L_{\mathrm{ins}}\)

The total HGST objective is:
$$
L_{\mathrm{total}} =
\lambda_{\mathrm{consistency}}L_{\mathrm{consistency}} +
\lambda_{\mathrm{style}}L_{\mathrm{style}} +
\lambda_{\mathrm{id}}L_{\mathrm{id}} +
\lambda_{\mathrm{illum}}L_{\mathrm{illum}} +
\lambda_{\mathrm{ins}}L_{\mathrm{ins}},
$$
with appendix weights
$$
\lambda_{\mathrm{consistency}} = 3,\quad
\lambda_{\mathrm{style}} = 18,\quad
\lambda_{\mathrm{id}} = 7,\quad
\lambda_{\mathrm{illum}} = 10^{-5},\quad
\lambda_{\mathrm{ins}} = 1.
$$

The paper’s stated rationale is sequential: geometry is learned once by 4DGS from content images; HGST produces structurally faithful, temporally consistent stylized supervision frames; and the Style Gaussian Representation applies per-Gaussian, depth- and time-aware modulation to yield fine, view-consistent style details without geometry drift.

## 5. Training protocol, rendering regime, and reproducibility

Training is staged sequentially. First, **4DGS geometry** is trained on multi-view content images \(I_{\mathrm{content}}\) using the 4DGS setup, with hyperparameters following Wu et al., CVPR’24. Second, **HGST** is trained on **COCO2014** content images and **WikiArt** style images using **\(256 \times 256\)** crops. Third, the **Style Gaussian** module is trained against HGST-generated stylized frames [2508.19243].

For stylization supervision, the Style4D objective is:
$$
L = \lVert \hat{I} - St(I) \rVert_1 + L_{\mathrm{tv}},
$$
where \(\hat{I}\) is the rendered image from Style4D and \(St(I)\) is the HGST stylized target, and \(L_{\mathrm{tv}}\) is total variation regularization.

Initialization uses the official 4DGS point cloud and downsamples it to approximately **3–4k Gaussians per scene**. The per-Gaussian MLPs are again specified as **two-layer** networks, also reported with **4 hidden units** and **4 outputs**, trained with a learning-rate schedule from **\(1\times 10^{-4}\)** to **\(1\times 10^{-5}\)** and **delay\_mult \(= 0.02\)**.

The main training logistics are explicit:

- **Batch size**: \(2\)
- **Iterations**: up to **14,000** per scene
- **Hardware**: **single NVIDIA A40 (48 GB)**
- **Training time**: approximately **2 hours per scene**
- **Peak memory**: approximately **10 GB**

Inference is performed at the native Neu3D resolution of **\(1352 \times 1014\)**. Although the method builds on real-time 4DGS rendering, the paper does **not** report frames per second for Style4D. Rendering complexity is described as scaling with the number of Gaussians and the cost of the per-Gaussian MLPs.

The project page is listed as the primary access point for the benchmark protocol, baseline, and links. The data are obtained under Neu3D’s license and terms. The setup is summarized as: Stage 1, train 4DGS on content frames; Stage 2, train HGST on COCO2014 and WikiArt with the specified loss weights; Stage 3, train per-Gaussian MLPs with \(L_1 + TV\) against HGST stylized supervision.

## 6. Quantitative performance, user studies, and limitations

Quantitative comparisons are reported on three representative scenes—`cook_spinach`, `flame_salmon_1`, and `sear_steak`—against **4DGS(AdaIN)**, **4DGS(AdaAttN)**, and **4DStyleGaussian**. Across these scenes, Style4D attains the best or tied-best results across most metrics, especially on **spatial fidelity** measured by **SSIM** and **LPIPS-to-content**, **aesthetics**, **CLIP-IQA+**, **MUSIQ**, and **DISTS** [2508.19243].

For `cook_spinach`, Style4D reports **UIQM 1.9290**, **CLIP-IQA+ 0.4437**, **MUSIQ 53.4681**, **Q-Align 3.2072**, **MUSIQ-PAQ2PIQ 65.8520**, **DISTS 0.0112**, **Warp 0.0058**, **DINO 0.9395**, **CKDN 0.2290**, **LPIPS-to-style 0.6866**, **SSIM 0.7771**, and **LPIPS-to-content 0.1834**. For `flame_salmon_1`, the corresponding values are **1.7529**, **0.3962**, **55.2302**, **3.6030**, **63.5178**, **0.0138**, **0.0067**, **0.9415**, **0.2354**, **0.7602**, **0.6963**, and **0.2704**. For `sear_steak`, Style4D reports **1.6818**, **0.4176**, **53.3443**, **2.8820**, **68.0488**, **0.0108**, **0.0066**, **0.9564**, **0.3239**, **0.7014**, **0.7503**, and **0.2146**, respectively. The paper notes a single clear exception: on `sear_steak`, **4DStyleGaussian** has the lowest warp loss (**0.0050**) while Style4D is second-best, though still best on DISTS.

The user studies reinforce the quantitative findings. In the **single-frame** study, voting percentages for **stylization quality / image quality** are **11.76% / 2.94%** for 4DGS(AdaIN), **14.70% / 8.82%** for 4DGS(AdaAttN), **11.76% / 17.64%** for 4DStyleGaussian, and **61.76% / 70.58%** for Style4D. In the **video** study, pairwise against Style4D, the percentages for **stylization quality / video quality / spatiotemporal consistency** are:
- **vs 4DStyleGaussian**: **23.52% / 30.12% / 44.11%** for 4DStyleGaussian versus **76.47% / 69.87% / 55.88%** for Style4D
- **vs 4DGS(AdaIN)**: **20.58% / 11.76% / 14.70%** versus **79.41% / 88.23% / 85.29%**
- **vs 4DGS(AdaAttN)**: **14.70% / 17.64% / 14.70%** versus **85.29% / 82.35% / 85.29%**

The ablations further isolate component behavior. For **2D video stylization temporal warp loss**, the appendix reports **AdaIN 0.045824**, **AdaAttN 0.031671**, **CCPL 0.013301**, **MCCNet 0.021858**, and **HGST 0.021068**. For **content preservation on 4D stylization**, Style4D exceeds the listed baselines on all three benchmarked scenes in **SSIM** and **LPIPS-to-content**. For **style loss**, measured as VGG Gram-matrix MSE to the style image, Style4D is lower than 4DStyleGaussian on `sear_steak` (**0.005687** vs **0.006816**), `cook_spinach` (**0.006123** vs **0.008329**), and `flame_salmon_1` (**0.006364** vs **0.007387**). An auxiliary reconstruction ablation shows that the style-aware Gaussian representation does not harm rendering quality, with PSNR improvements over 4DGS on all six original scenes, including `flame_steak` (**32.85** vs **32.02**) and `cut_roasted_beef` (**32.81** vs **32.55**).

The paper identifies several limitations. **Computational cost** remains nontrivial because of multi-stage training and per-Gaussian MLP overhead; the reported cost is approximately two hours per scene on an A40 with about 10 GB peak memory. **Style granularity** is limited to a **fixed style per scene**; rapid style switching and region-specific stylization are not supported. The paper also notes that, although temporal stability is strong, **very fast motions or extremely complex occlusions remain challenging in general**. The stated future direction is to improve stylization quality further and enable broader, user-controllable style manipulation for interactive applications.

Taken together, the benchmark’s design emphasizes standardized evaluation of dynamic 3D stylization through simultaneous measurement of geometry-faithful rendering, temporal stability, and view consistency. A plausible implication is that Style4D-Bench’s principal contribution lies not only in reporting a strong baseline, but in formalizing a metric and data regime under which future 4D stylization methods can be compared on the same spatiotemporal and multiview criteria.

Source: https://www.emergentmind.com/topics/style4d-bench