Due to supply constraints, replacing damaged automotive components can be expensive and time-consuming, while replacing the entire component can be wasteful when much of it remains usable. This paper presents a pipeline to investigate whether missing geometry can be reconstructed as a repair patch using paired observations of the intact models and synthetically damaged components. Noise-free color, depth, and silhouette images were produced using the camera viewpoints. Depth differences were then used to identify the potential missing regions, while multi-view voxel reconstruction was used to estimate the missing portion. A pretrained DINOv2 model was also added to validate missing candidates and determine whether pretrained visual features could improve reconstruction. We tested the pipeline on 20 automotive CAD models, each with a synthetically removed portion. Each model was tested under four different conditions using two reconstruction methods for a total of 160 trials. The predictions were compared with the missing geometry of the unchanged ground-truth using predefined thresholds of the voxel intersection-over-union over 0.85 and the absolute relative volume error less than 15%. Of the 160 experimental trials, 109 of them satisfied both thresholds. Geometry-only reconstruction passed 16 of 20 components with eight views but only 10 of 20 with four adjacent views. Further analysis showed that 24 out of the 37 failed reconstructions could be categorized as unsuccessful because extra material was incorrectly added to areas that should have been empty. The implementation of DINOv2 into the pipeline did not improve the overall reconstruction accuracy and made prediction generation slower. For 13 verified tests, printing only the repaired sections instead of the entire parts reduced filament use by 93.08% and printing time by 89.39% based on the median results.
Figures & tables
Figure 1: Overview of the CAD-guided reconstruction pipeline for generating the repair patch
Figure 2: The 20 automotive CAD components used in the evaluation. Green identifies the synthetically removed region for each component.
Figure 3: Comparison of exhaust flange reconstruction under different camera conditions. (a) Eight viewpoints (IoU = 0.898). (b) Four adjacent viewpoints (IoU = 0.369). Red indicates extra material; blue indicates missed material.
Views
Method
Pass
Fail
Refused
8
Geometry
16
3
1
8
DINOv2
16
3
1
4 spread
Geometry
13
6
1
4 spread
DINOv2
12
4
4
4 adjacent
Geometry
10
8
2
4 adjacent
DINOv2
10
7
3
Table 1: Reconstruction outcomes across all 160 trials.
Component
8 views
4 spread
4 adjacent
8 darkened
Ford parcel-shelf clip
94
92
93
94
Hyundai armrest latch
93
86
92
93
Kia/Hyundai vent clip
91
89
90
91
Mazda tow-hook cap
93
89
92
93
Renault lever cap
97
93
97
97
Tesla cabin-filter clip
91
87
88
91
Table 2: Per-component voxel IoU (%) for geometry-only reconstruction across camera conditions.
Figure 4: Comparison of geometry-only and DINOv2-assisted reconstruction. (a) Passing trials. (b) Accuracy differences among paired repairs. (c) Median prediction-stage runtime.
Measurement
Reduction
Solid volume
95.60%
Estimated filament use
93.08%
Estimated printing time
89.39%
Table 3: Median reductions for 13 verified paired slicing comparisons.
High-fidelity vehicle assets are essential for controllable traffic scene generation, particularly for synthesizing rare and safety-critical long-tail scenarios. However, reconstructing a reusable vehicle representation from in-the-wild onboard images remains challenging for two reasons. First, image-to-3D generation methods generally produce models without reliable metric scale. Second, onboard cameras usually observe only one side of a target vehicle, making conventional multi-view reconstruction incomplete on unobserved regions. To solve these problems, we propose a feed-forward vehicle asset reconstruction method, which leverages two complementary priors to reconstruct 3D Gaussian representations for vehicles using sparse one-sided observations. To achieve metric-scale reconstruction, a visual foundation model is first utilized to serve as a visual prior for Gaussian initialization. The Gaussian attributes are then estimated by a learnable encoder-decoder module. A symmetry-aware cloning strategy is presented to complete the unobserved side directly in Gaussian space, which exploits the bilateral structure of vehicles as a geometric prior. Experiments on the public dataset demonstrate that the proposed method significantly outperforms existing approaches in both vehicle asset completeness and geometric accuracy.
Jinyu Miao, Jiusi Li, Yifei He +4
School of Vehicle and Mobility and the State Key Laboratory of Intelligent Green Vehicle and Mobility, Tsinghua University, Beijing, China
Per-scene optimization methods such as 3D Gaussian Splatting provide state-of-the-art novel view synthesis quality but extrapolate poorly to under-observed areas. Methods that leverage generative priors to correct artifacts in these areas hold promise but currently suffer from two shortcomings. The first is scalability, as existing methods use image diffusion models or bidirectional video models that are limited in the number of views they can generate in a single pass (and thus require a costly iterative distillation process for consistency). The second is quality itself, as generators used in prior work tend to produce outputs that are inconsistent with existing scene content and fail entirely in completely unobserved regions. To solve these, we propose a two-stage pipeline that leverages two key insights. First, we train a powerful bidirectional generative model with a novel opacity mixing strategy that encourages consistency with existing observations while retaining the model's ability to extrapolate novel content in unseen areas. Second, we distill it into a causal auto-regressive model that generates hundreds of frames in a single pass. This model can directly produce novel views or serve as pseudo-supervision to improve the underlying 3D representation in a simple and highly efficient manner. We evaluate our method extensively and demonstrate that it can generate plausible reconstructions in scenarios where existing approaches fail completely. When measured on commonly benchmarked datasets, we outperform all existing baselines by a wide margin, exceeding prior state-of-the-art methods by 1-3 dB PSNR.
Riccardo de Lutio, Tobias Fischer, Yen-Yu Chang +7
NVIDIA, USA · NVIDIA, Switzerland · ETHZ, Switzerland +4
CAD practitioners often work with proprietary parts that they are unwilling to share with external services. For individuals and small companies, bringing this work in-house also requires computational resources they can afford. We introduce CADMAGEDDON, an open-source agent that reconstructs editable CAD from meshes and can run locally on single-GPU systems and current-generation laptops. The method decomposes this complex task into a chain of simple visual decisions about what to construct and what to correct. Geometric tools support precise construction, allowing the agent to focus on visual reasoning. Using the mid-sized Qwen3.8-27B model, CADMAGEDDON outperforms prior approaches, including systems based on frontier models, in reconstruction quality on all the main benchmarks. CADMAGEDDON integrates with FreeCAD, bringing its reconstructions into practitioners' existing CAD workflows. Code: https://github.com/kulibinai/cadmageddon
Oleg Pavlov, Nikita Gavrilov, Soslan Kabisov +4
Lomonosov Moscow State University · AXXX, Moscow, Russia · Innopolis University +1