CADReasoner: Iterative Program Editing for CAD Reverse Engineering

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kabisov, Soslan, Kirichuk, Vsevolod, Volkov, Andrey, Savrasov, Gennadii, Barannikov, Marina, Konushin, Anton, Kuznetsov, Andrey, Zhemchuzhnikov, Dmitrii
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911558064406528
author Kabisov, Soslan
Kirichuk, Vsevolod
Volkov, Andrey
Savrasov, Gennadii
Barannikov, Marina
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
author_facet Kabisov, Soslan
Kirichuk, Vsevolod
Volkov, Andrey
Savrasov, Gennadii
Barannikov, Marina
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
contents Computer-Aided Design (CAD) powers modern engineering, yet producing high-quality parts still demands substantial expert effort. Many AI systems tackle CAD reverse engineering, but most are single-pass and miss fine geometric details. In contrast, human engineers compare the input shape with the reconstruction and iteratively modify the design based on remaining discrepancies. Agent-based methods mimic this loop with frozen VLMs, but weak 3D grounding of current foundation models limits reliability and efficiency. We introduce CADReasoner, a model trained to iteratively refine its prediction using geometric discrepancy between the input and the predicted shape. The model outputs a runnable CadQuery Python program whose rendered mesh is fed back at the next step. CADReasoner fuses multi-view renders and point clouds as complementary modalities. To bridge the realism gap, we propose a scan-simulation protocol applied during both training and evaluation. Across DeepCAD, Fusion 360, and MCB benchmarks, CADReasoner attains state-of-the-art results on clean and scan-sim tracks.
format Preprint
id arxiv_https___arxiv_org_abs_2603_29847
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CADReasoner: Iterative Program Editing for CAD Reverse Engineering
Kabisov, Soslan
Kirichuk, Vsevolod
Volkov, Andrey
Savrasov, Gennadii
Barannikov, Marina
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
Graphics
Computer Vision and Pattern Recognition
Human-Computer Interaction
Computer-Aided Design (CAD) powers modern engineering, yet producing high-quality parts still demands substantial expert effort. Many AI systems tackle CAD reverse engineering, but most are single-pass and miss fine geometric details. In contrast, human engineers compare the input shape with the reconstruction and iteratively modify the design based on remaining discrepancies. Agent-based methods mimic this loop with frozen VLMs, but weak 3D grounding of current foundation models limits reliability and efficiency. We introduce CADReasoner, a model trained to iteratively refine its prediction using geometric discrepancy between the input and the predicted shape. The model outputs a runnable CadQuery Python program whose rendered mesh is fed back at the next step. CADReasoner fuses multi-view renders and point clouds as complementary modalities. To bridge the realism gap, we propose a scan-simulation protocol applied during both training and evaluation. Across DeepCAD, Fusion 360, and MCB benchmarks, CADReasoner attains state-of-the-art results on clean and scan-sim tracks.
title CADReasoner: Iterative Program Editing for CAD Reverse Engineering
topic Graphics
Computer Vision and Pattern Recognition
Human-Computer Interaction
url https://arxiv.org/abs/2603.29847