«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

期数:

2025年07期

页码:

16-23

栏目:

软件技术与工程

出版日期:

2025-07-10

文章信息/Info

Title:

Test Data Generation for Object Detection Based on Automated Semantic Editing

文章编号:

1673-629X(2025)07-0016-08

作者:

陈皓明1; 桂智明1; 刘艳芳2; 范鑫鑫3; 路云峰4

1. 北京工业大学 计算机学院,北京 100124;

2. 北京航空航天大学 计算机学院,北京 100083;

3. 中国科学院 计算技术研究所,北京 100190;

4. 北京航空航天大学 可靠性与系统工程学院,北京 100088

Author(s):

CHEN Hao-ming1; GUI Zhi-ming1; LIU Yan-fang2; FAN Xin-xin3; LU Yun-feng4

1. School of Computer Science and Technology,Beijing University of Technology,Beijing 100124,China;

2. School of Computer Science and Engineering,Beihang University,Beijing 100083,China;

3. Institute of Computing Technology,Chinese Academy of Sciences,Beijing 100190,China;

4. School of Reliability and Systems Engineering,Beihang University,Beijing 100088,China

关键词:

目标检测; 语义编辑; 测试数据生成; 深度神经网络; 图像生成

Keywords:

object detection; semantic editing; test data generation; deep neural networks; image generation

分类号:

TP391

DOI:

10.20165/j.cnki.ISSN1673-629X.2025.0052

摘要:

目标检测系统的测试数据生成对评估模型性能和发现潜在缺陷至关重要。 现有方法在生成数据的多样性和真实性方面仍存在局限。 该文提出了一种基于自动语义编辑的目标检测测试数据生成方法 SemaGen,通过构建高质量语义对象库并结合自动化语义编辑策略,实现对图像的插入、删除和替换等高级语义操作。具体而言,该方法首先通过多重筛选机制构建语义对象库,确保对象的语义完整性和场景适应性;其次,利用场景复杂度量化模型,综合考虑背景占比、实例数量和空间分布等因素,实现编辑策略的自适应选择;最后,提出基于对象重要性的替换策略、迭代式删除方法以及考虑语义相似度的智能插入机制,确保生成图像的真实性和多样性。 实验结果表明,SemaGen 在三种对象操作任务上显著优于现有方法,生成的图像质量更高,FID 得分更优,证实了该方法在生成数据质量方面的优越性。 在目标检测模型测试中,SemaGen 成功发现 YOLO v11、SSD 和 Mask R-CNN 等主流检测器在复杂场景下的性能缺陷,为目标检测测试用例生成提供了新的思路和工具。

Abstract:

Test data generation for object detection systems is crucial for evaluating model performance and identifying potential defects.Existing methods still have limitations in generating diverse and realistic data. We present SemaGen,a test data generation method for object detection based on automated semantic editing, which achieves advanced semantic operations such as insertion, deletion, and replacement through constructing high - quality semantic object libraries and combining automated editing strategies. Specifically, the proposed method first constructs a semantic object library through multiple screening mechanisms to ensure object semantic integrity and scene adaptability. Secondly, it utilizes a scene complexity quantification model that comprehensively considers background ratio, instance quantity,and spatial distribution to achieve adaptive selection of editing strategies. Finally,it proposes an object importance-based replacement strategy,an iterative deletion method,and an intelligent insertion mechanism considering semantic similarity to ensure the authenticity and diversity of generated images. The experimental results show that SemaGen significantly outperforms the existing methods on the three object manipulation tasks,generates higher quality images with better FID scores,and confirms its superiority in gen-erating data quality. In object detection model testing,SemaGen successfully identifies performance deficiencies of mainstream detectors such as YOLO v11,SSD,and Mask R-CNN in complex scenarios,providing new insights and tools for generating object detection test cases.

常用功能

更新日期/Last Update: 2025-07-10