Cinematic Pre-Visualization System: Automated 3D Scene Generation from Natural Language Descriptions Using NLP and JSON-Based Scene Representation

Text-to-hierarchical three-dimensional scene generation: a new approach for layered three-dimensional modeling from natural language

Rapid advancements in generative models have enabled substantial progress in text-to-3D scene synthesis. However, existing approaches often lack hierarchical structure and geometric consistency, limiting their use in complex, editable scene creation. To address this issue, an end-to-end hierarchical framework for text-to-3D scene generation is introduced. The main contribution of this study lies not in advancing a single generative model but in the design of a framework that synergistically integrates state-of-the-art components for video synthesis and mesh reconstruction. By decoupling scene semantics from dynamic camera trajectory control, the proposed system efficiently produces structured object-level editable three-dimensional scenes from natural language. Comprehensive experiments showed that the proposed integrated approach achieved superior scene quality, consistency, and editing flexibility compared with existing methods, offering a powerful solution for applications, such as virtual reality and digital twins.

BIMScript: Material-Aware Structured Scene Programs for BIM Ingestion

Structured-language models such as SceneScript reconstruct a scene as a short program of parametric commands, an inherently editable and semantically explicit representation. We ask three questions that stand between such models and their most compelling application, automated ingestion of existing buildings into BIM tools, studied here on synthetic scans: \emph{what} is the scene made of, \emph{how fast} can it be produced, and \emph{exactly where} is each element. BIMScript answers all three within one grammar. First, we extend the layout language with per-element \emph{material} and \emph{condition} attributes, supervised by a vision-language-model material-passport corpus we build over 100k synthetic scenes (1.9M pseudo-labeled elements), and route image appearance to the material tokens through a lifted-feature point encoder. Second, we show that autoregressive decoding of these programs is dominated not by compute but by kernel-launch and host-synchronization overhead, and remove it with an output-exact CUDA-graph decoder (1.9 vs 6.4\,ms/step, $3.4\times$) plus a grammar-parallel, tolerance-verified draft-and-verify scheme that exploits the deterministic entity schema. Third, we address the model's 5cm token-grid granularity with training-free geometric snapping and a hybrid discrete--continuous decoder head that regresses a sub-bin offset, and measure how much of the residual error each recovers. Because each command maps one-to-one onto a native Revit object, we validate direct ingestion into a BIM authoring tool end to end with a working add-in and its IFC4 export, and the same program's language form is designed to support LLM-driven, sustainability-aware reasoning over the built asset.

Object-Level Access to SVG Scenes: A Benchmark and Raster-Grounded Method for Text-Guided Vector Object Extraction

LLM-Supervised Point Cloud Processing: From Unsupervised 3D Scene-Graph Generation to Interactive Scene Manipulation

Abstract. We demonstrate an end-to-end pipeline for 3D scene understanding which integrates unsupervised graph-based point cloud segmentation with LLM-enabled spatial reasoning and editing. A point cloud is segmented into a SemanticPatch decomposition (stage 1), labeled using a zero-shot vision-language model (stage 2; SAMv2, CLIP), encoded into a scene graph in the latent space (stage 3) capturing geometry, topology, and constraints, and finally manipulated by an LLM-based agent (stage 4) to execute a specified editing task. The LLM agent can be instructed by natural language input to reason about a scene graph and a point cloud, compute a geometric transformation for the input point cloud, and check its own output against a set of constraints (e.g. ADA-compliance). We validate our approach on three different point clouds: a classroom (Leica RTC360, 1.3 M points), a construction site (NavVis VLX mobile scanner, 4.4M points), and the Paris-Lille-3D benchmark. Our segmentation approach scores 97–99% on the fitness score and 92–99% on the F1-score across all three benchmarks. Our LLM agent solves reconfiguration tasks in 1–10 min, achieving a 100% constraint-satisfaction rate and outperforming a human annotator.

CinemaTraj: Composing Atomic Camera Trajectories for 3D Scenes with LLM Agents

Automatically generating cinematically expressive camera trajectories through 3D scenes from natural language descriptions is a challenging task of high practical value, with applications ranging from real-estate advertising to virtual tour creation. Existing methods either lack true 3D spatial awareness by relying on 2D image priors, or treat trajectory generation as a geometric path planning problem divorced from cinematographic semantics. We present CinemaTraj, a framework that reframes camera trajectory planning as a language-grounded spatial reasoning problem. Given a set of RGB-D images and a user prompt, CinemaTraj equips an LLM agent with a structured 3D scene graph: the agent decomposes the prompt into a sequence of atomic cinematographic movements (dolly, orbit, crane, pan, tilt, zoom, arc). Each movement is instantiated via a novel parametric trajectory representation that is both cinematographically expressive and optimizable for collision avoidance. The scene graph acts as a structured spatial prior, grounding the agent's reasoning in accurate geometric and semantic knowledge of the environment. CinemaTraj further generates synchronized voiceover and subtitles aligned with camera motion, producing narrated cinematic video outputs. We evaluate CinemaTraj on real-world ScanNet++ environments, and show that it produces prompt-faithful, collision-free trajectories with high cinematographic quality, outperforming existing approaches on prompt alignment, trajectory quality, and safety metrics.

From text to immersive 3D: an end-to-end pipeline for automated retrieval-augmented virtual room generation

This paper introduces a retrieval-augmented Text-to-WebXR pipeline that converts free text into interactive, browser-based 3D environments. Leveraging embedding-based retrieval, cross-encoder reranking, and a structured scene schema with constraint-aware placement, the approach generated one validated room description for each of the 229 firms. Expert evaluation indicates consistently high quality (M=4.46/5), with near-perfect scores in consistency (4.99) and completeness (4.88). A user study with 204 participants shows that 79.6% perceive 3D environments as a meaningful enhancement to textual profiles; participants with less prior 3D experience found the representations more helpful, indicating a novelty effect. These results demonstrate that immersive WebXR representations substantially strengthen user engagement and understanding, offering a scalable pathway for employer branding and digital recruitment, and point to interactivity as a key avenue for future work.