| 引用本文: | 肖欣怡,柳林,郭茜,罗玉川,王勇军,陈荣茂,黄俊杰,刘天瑞,付绍静.恶意敌手环境下的隐私保护目标检测.软件学报,2026,37(6):2411-2430 |
| |
|
| |
|
|
| 本文已被：浏览 573次 下载 269次 |
码上扫一扫！ |
|
|
| 恶意敌手环境下的隐私保护目标检测 |
|
肖欣怡1, 柳林1,2, 郭茜1, 罗玉川1, 王勇军1, 陈荣茂1,2, 黄俊杰1, 刘天瑞1, 付绍静1,2
|
|
1.国防科技大学 计算机学院, 湖南 长沙 410073;2.国防科技大学 密码中心, 湖南 长沙 410073
|
|
| 摘要: |
| 图像处理任务正快速向云端和多方协同环境迁移, 而云服务器上直接处理明文图像数据, 极易泄露图像中的敏感信息, 且难以抵御篡改等恶意攻击, 无法保证数据完整性和服务可靠性. 在此背景下, 提出一种面向恶意敌手环境的目标检测推理方案——MalOD, 实现针对恶意敌手环境的安全目标检测. MalOD通过构建加密的特征金字塔网络 (secure feature pyramid network, SecFPN)实现密文图像的多级特征提取, 并基于多层次密文特征设计安全区域提议网络(secure region proposal network, SecRPN)和兴趣区域安全对齐(secure region of interest align, SecRoIA)模块从而完成安全目标检测. 具体来说, 借助复制秘密共享(replicated secret sharing, RSS)技术, 设计一系列安全计算原语, 包括安全向上取整函数、安全双线性插值和安全最近邻插值, 为SecFPN、SecRPN、SecRoIA等模块提供底层支撑, 确保恶意敌手环境下检测流程的高效与准确. 此外, 证明MalOD的正确性和安全性, 并在COCO 2017和Pascal VOC 2012数据集上进行性能评估. 实验结果表明, 在满足严格安全要求的同时, MalOD实现较高的目标检测精度. 特别地, 当目标检测的交并比阈值为0.5时, 其在COCO子集上平均精度仅比明文检测下降0.113. 为恶意环境下的隐私保护图像处理提供了理论和实践支持, 尤其适用于不可信的云计算和多方协作场景中. |
| 关键词: 隐私保护 目标检测 恶意敌手 安全多方计算 |
| DOI：10.13328/j.cnki.jos.007648 |
| 分类号:TP309 |
| 基金项目:国家自然科学基金 (62472431); 湖南省自然科学基金 (2023JJ30640) |
|
| Privacy-preserving Object Detection Under Malicious Adversaries |
|
XIAO Xin-Yi1, LIU Lin1,2, GUO Qian1, LUO Yu-Chuan1, WANG Yong-Jun1, CHEN Rong-Mao1,2, HUANG Jun-Jie1, LIU Tian-Rui1, FU Shao-Jing1,2
|
|
1.College of Computer Science and Technology, National University of Defense Technology, Changsha 410073, China;2.Center for Cryptologic Research, National University of Defense Technology, Changsha 410073, China
|
| Abstract: |
| Image processing tasks are rapidly migrating to cloud and multi-party collaborative environments. However, directly processing plaintext image data on cloud servers easily leads to the leakage of sensitive information in images and is difficult to resist malicious attacks such as tampering, thus failing to guarantee data integrity and service reliability. To address these challenges, this study proposes MalOD, an object detection inference framework for environments with malicious adversaries. MalOD is a framework to achieve secure object detection under malicious adversaries. MalOD constructs an encrypted feature pyramid network (SecFPN) to perform multi-level feature extraction on encrypted images. Based on these multi-level cipher text features, a secure region proposal network (SecRPN) and a secure region of interest align (SecRoIA) module are designed to achieve secure object detection. By leveraging replicated secret sharing (RSS), a series of secure computation primitives are designed, including a secure ceiling function, secure bilinear interpolation, and secure nearest-neighbor interpolation. These primitives provide the underlying support for SecFPN, SecRPN, and SecRoIA, ensuring the efficiency and accuracy of the detection process under malicious adversaries. The correctness and security of MalOD are proved, and its performance is evaluated on the COCO 2017 and Pascal VOC 2012 datasets. Experimental results show that MalOD achieves high object detection accuracy while meeting strict security requirements. In particular, when the intersection over union (IoU) threshold is 0.5, the average precision on the COCO subset decreases by only 0.113 compared with plaintext detection. This study provides theoretical and practical support for privacy-preserving image processing under malicious environments and is particularly suitable for untrusted cloud computing and multi-party collaboration scenarios. |
| Key words: privacy-preserving object detection malicious adversary secure multi-party computation |
|
|
|
|
