Create README.md

README.md

@@ -0,0 +1,59 @@
+# Img2CADSeq
+
+**[SIGGRAPH 2026] Img2CADSeq: Image-to-CAD Generation via Sequence-Based Diffusion**
+
+[![Project Page](https://img.shields.io/badge/Project-Page-blue)](https://rilpraa0110.github.io/Img2CADSeq/)
+[![arXiv](https://img.shields.io/badge/arXiv-2605.13293-b31b1b.svg)](https://arxiv.org/abs/2605.13293)
+[![GitHub](https://img.shields.io/badge/GitHub-Code-black)](https://github.com/Rilpraa0110/Img2CADSeq)
+
+Official repository for **Img2CADSeq**, a pipeline that outputs standardized STEP files.
+
+---
+
+## 📖 Abstract
+
+Boundary Representation (BRep) is the standard format for Computer-Aided Design (CAD), yet reconstructing high-quality BReps from single-view images remains challenging due to the complexity of topological constraints and operation sequences. 
+
+We present **Img2CADSeq**, a multi-stage pipeline that overcomes these limitations by encoding CAD sequences into a three-level hierarchical codebook. Guided by an importance prioritization, this strategy values profiles over details, compressing long sequences into a stable discrete latent space. To bridge the modality gap, we leverage a coarse-to-fine point cloud intermediate, aligning 2D visual features with 3D CAD sequences via contrastive learning to condition a VQ-Diffusion model. 
+
+Supported by newly introduced **CAD-220K** and **PrintCAD** datasets, our approach ensures robust industrial domain adaptation. Extensive experiments demonstrate that Img2CADSeq significantly outperforms state-of-the-art methods, producing standard STEP files that can be directly used in commercial CAD software.
+
+---
+
+## 📊 Datasets
+
+To train and evaluate our framework, we introduce two distinct data types: curated synthetic models and real-world captured objects. Both datasets are exclusively utilized to fine-tune our image-to-point-cloud module, ensuring that the point cloud embeddings align with the CAD latent space.
+
+### CAD-220K (Synthetic)
+To support the point cloud lifting stage, we utilize CAD-220K, a curated subset of the ABC dataset filtered by surface count. Observing that models with 11–50 faces constitute the vast majority (339,489 in total), we proportionally downsample the data to establish balanced complexity tiers:
+* **40K** models (1–10 faces)
+* **120K** models (11–50 faces)
+* **30K** models (51–100 faces)
+* **30K** models (>100 faces)
+
+For these 220K models, we generate the corresponding STLs, point clouds, and four-view rendered images. 
+> 📁 **Repository Files:** The filtered surface count lists and selected IDs for the CAD-220K dataset are available in this repository as `all_surfs_result.csv` and `all_surfs_result.json`.
+
+### PrintCAD (Real-world)
+To explore sim-to-real translation, we introduce PrintCAD, a collection of over 2,000 3D-printed solids. For each model, we systematically capture four views under real-world lighting. These objects exhibit manufacturing artifacts and texture noise, offering an evaluation of model robustness beyond synthetic renders. 
+> 📁 **Repository Files:** The complete PrintCAD dataset is available for download in this repository as `PrintCAD.zip`.
+
+---
+
+## ⚙️ Workflow Overview
+
+At the base level, the **Curve-Cluster** parameterizes geometric primitives, which form closed loops in the **Sketch-Patch**. These loops are then lifted into 3D space via a normal vector and origin to perform extrusion and Boolean operations, resulting in an **Extrude-Block**. Multiple blocks are finally assembled to yield the target solid. This process mirrors the construction history of standard CAD workflows, preserving human design intent.
+
+---
+
+## 📝 Citation
+
+If you find our work or datasets useful in your research, please consider citing:
+
+```bibtex
+@article{tan2026img2cadseq,
+  title={Img2CADSeq: Image-to-CAD Generation via Sequence-Based Diffusion},
+  author={Tan, Shiyu and Zhao, Zixuan and Gao, Hao and Chen, Zhiheng and Yin, Xiaolong and Shen, Enya},
+  journal={arXiv preprint arXiv:2605.13293},
+  year={2026}
+}