Image tampering localization and detection network under brightness-contrast disturbances

doi:10.11772/j.issn.1001-9081.2025050655

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (6): 1893-1903.DOI: 10.11772/j.issn.1001-9081.2025050655

• Cyber security • Previous Articles

Image tampering localization and detection network under brightness-contrast disturbances

Xiaoqin YU¹, Wuyang SHAN¹(), Junying QIU², Yu LIN³, Ronghao YANG⁴, Mao TIAN⁵

^1.College of Computer Science and Network Security （Model Software College），Chengdu University of Technology，Chengdu Sichuan 610059，China
^2.College of Fashion and Design Art，Sichuan Normal University，Chengdu Sichuan 610066，China
^3.Business School，Chengdu University of Technology，Chengdu Sichuan 610059，China
^4.College of Earth and Planetary Sciences，Chengdu University of Technology，Chengdu Sichuan 610059，China
^5.School of Computer Science and Technology，Chongqing University of Posts and Telecommunications，Chongqing 400065，China

Received:2025-06-12 Revised:2025-09-08 Accepted:2025-09-19 Online:2025-09-25 Published:2026-06-10
Contact: Wuyang SHAN
About author:YU Xiaoqin， born in 2003， M. S. candidate. Her research interests include image forensics， deep learning.
QIU Junying， born in 1990， M. S.， lecturer. Her research interests include intelligent processing， analysis and security of digital art media.
LIN Yu， born in 1973， Ph. D.， professor. His research interests include financial risk management， artificial intelligence， intelligent risk early warning.
YANG Ronghao， born in 1978， Ph. D.， professor. His research interests include computer vision， intelligent processing of remote sensing data.
TIAN Mao， born in 1988， Ph. D.， lecturer. His research interests include stereo matching， point cloud and image fusion， deep learning.
First author contact:SHAN Wuyang， born in 1988， Ph. D.， associate professor. His research interests include multimedia forensics， computer vision， data hiding.
Supported by:
National Natural Science Foundation of China(42001417);Humanities and Social Sciences Planning Fund of Ministry of Education(2024XJAZH005)

亮度对比度扰动下的图像篡改定位检测网络

喻小芹¹, 单武扬¹(), 邱骏颖², 林宇³, 杨容浩⁴, 田茂⁵

^1.成都理工大学计算机与网络安全学院（示范性软件学院），成都 610059
^2.四川师范大学服装与设计艺术学院，成都 610066
^3.成都理工大学商学院，成都 610059
^4.成都理工大学地球与行星科学学院，成都 610059
^5.重庆邮电大学计算机科学与技术学院，重庆 400065

通讯作者: 单武扬
作者简介:喻小芹（2003—），女，四川仪陇人，硕士研究生，CCF会员，主要研究方向：图像取证、深度学习
邱骏颖（1990—），女，四川成都人，讲师，硕士，主要研究方向：数字艺术媒体的智能处理、分析和安全
林宇（1973—），男，四川仪陇人，教授，博士，主要研究方向：金融风险管理、人工智能、智能风险预警
杨容浩（1978—），男，湖北天门人，教授，博士，主要研究方向：计算机视觉、遥感数据的智能处理
田茂（1988—），男，四川德阳人，讲师，博士，CCF会员，主要研究方向：立体匹配、点云和影像融合、深度学习。
第一联系人：单武扬（1988—），男，湖南岳阳人，副教授，博士，主要研究方向：多媒体取证、计算机视觉、数据隐藏
基金资助:
国家自然科学基金资助项目(42001417);教育部人文社会科学规划基金资助项目(2024XJAZH005)

Abstract

Abstract:

Digital image tampering detection is critically important in the fields such as digital forensics and media content verification. However， in real-world applications， the tampered images are often post-processed in brightness and contrast， which will weaken tampering traces and degrade performance of the existing algorithms. To address this challenge， a restoration-assisted image tampering detection network ReConWave-Net was proposed. The network was consisted of two key modules： a classification-guided image restoration module was used to perform targeted restoration of images based on the categories of image disturbances， thereby reducing the impact of brightness and contrast disturbances； and a tampering localization module was used to strengthen the feature expression and localization ability of the tampered regions through multi-scale wavelet features and contrastive learning mechanism. The proposed network was evaluated on multiple datasets under various brightness and contrast disturbances. In terms of restoration quality， compared with the unrestored post-processed images， the proposed method increased the average Peak Signal-to-Noise Ratio （PSNR） in tampered regions from 10.86 dB to 31.57 dB， and improved the average Structural SIMilarity index （SSIM） from 0.40 to 0.92； in terms of detection performance， under typical disturbances， the network had the F1 score of 0.730 and an Intersection over Union （IoU） of 0.653. It can be seen that combining targeted restoration with detection can enhance the robustness of tampering localization of post-processed images significantly.

Key words: image forensics, image restoration, tampering detection, contrastive learning, wavelet feature enhancement

摘要：

数字图像篡改检测在数字取证和媒体内容验证等领域具有重要意义。然而，实际应用中篡改图像经常经历亮度和对比度等后处理操作，这会削弱篡改痕迹并降低现有算法的检测性能。针对这一问题，提出一种恢复辅助的图像篡改定位检测网络ReConWave-Net。该网络包含2个关键模块：分类引导的图像恢复模块用于根据图像扰动类别针对性地恢复图像，以减弱亮度和对比度扰动的影响；篡改定位模块则通过多尺度小波特征和对比学习机制，增强篡改区域的特征表达和定位能力。在多个数据集上及若干亮度和对比度扰动下评估所提网络的结果表明，在篡改区域恢复质量方面，相较于未恢复的后处理图像，所提网络将平均峰值信噪比（PSNR）由10.86 dB提高至31.57 dB，将平均结构相似性指标（SSIM）由0.40提高至0.92；在检测性能方面，典型扰动下的F1分数为0.730，而交并比（IoU）为0.653。可见，将针对性恢复与检测相结合可显著提升对后处理图像的篡改定位鲁棒性。

关键词: 图像取证, 图像恢复, 篡改检测, 对比学习, 小波特征增强

CLC Number:

TP751

Xiaoqin YU, Wuyang SHAN, Junying QIU, Yu LIN, Ronghao YANG, Mao TIAN. Image tampering localization and detection network under brightness-contrast disturbances[J]. Journal of Computer Applications, 2026, 46(6): 1893-1903.

喻小芹, 单武扬, 邱骏颖, 林宇, 杨容浩, 田茂. 亮度对比度扰动下的图像篡改定位检测网络[J]. 《计算机应用》唯一官方网站, 2026, 46(6): 1893-1903.

Figures/Tables 16

Fig. 1 Impact of post-processing on image tampering localization methods

Fig. 2 Network architecture of ReConWave-Net

Fig. 3 Local window attention and lightweight CNN replacement in Retinexformer

Fig.4 Comparison examples of weak artifact detection before and after restoration

Fig. 5 Structure of wavelet domain feature enhancement network

Fig. 6 Using flow of pixel-level contrast loss

Fig. 7 Complexity comparison of different detection methods

Tab. 1 Image restoration quality evaluation under different forgery types and brightness-contrast disturbances

篡改类型	方法	α=0.5		α=-0.5		β=0.7		β=2.0		平均值
篡改类型	方法	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM	PSNR/dB	SSIM
复制-移动	未恢复	7.26	0.47	8.16	0.18	13.27	0.51	12.44	0.66	10.28	0.46
	Retinexformer	14.21	0.56	23.12	0.80	26.40	0.82	15.23	0.74	19.74	0.73
	MSEC-Net	25.21	0.78	15.23	0.39	19.34	0.62	26.76	0.81	21.64	0.65
	本文恢复方法	31.02	0.93	29.39	0.90	32.90	0.93	31.33	0.93	31.16	0.92
拼接	未恢复	7.30	0.39	8.48	0.18	12.40	0.49	13.10	0.59	10.32	0.41
	Retinexformer	13.96	0.61	23.58	0.81	25.71	0.79	14.35	0.73	19.40	0.74
	MSEC-Net	25.41	0.80	14.27	0.44	20.00	0.59	26.12	0.78	21.45	0.65
	本文恢复方法	31.54	0.93	30.13	0.91	31.50	0.92	32.00	0.93	31.29	0.92
删除	未恢复	6.88	0.36	8.91	0.12	15.30	0.45	16.80	0.39	11.97	0.33
	Retinexformer	13.82	0.56	22.98	0.78	26.62	0.78	14.81	0.73	19.56	0.71
	MSEC-Net	25.12	0.81	14.63	0.39	19.52	0.57	26.98	0.78	21.56	0.64
	本文恢复方法	31.14	0.92	30.40	0.90	33.20	0.93	34.33	0.94	32.27	0.92

Tab. 2 Quantitative comparison of detection results of brightness processed images using F1 and IoU as criteria

数据集	方法	—		$α = 0.2$		$α = 0.5$		$α = - 0.2$		$α = - 0.5$		平均值
数据集	方法	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU
CASIA	MVSS-Net	0.45	0.40	0.40	0.32	0.35	0.29	0.35	0.26	0.25	0.19	0.36	0.29
	PSCC-Net	0.46	0.41	0.42	0.35	0.25	0.18	0.28	0.21	0.22	0.16	0.33	0.26
	TruFor	0.71	0.64	0.47	0.40	0.33	0.27	0.45	0.38	0.35	0.28	0.46	0.39
	MGQFormer	0.68	0.61	0.44	0.37	0.30	0.24	0.42	0.35	0.32	0.25	0.43	0.36
	SAFIRE	0.30	0.24	0.26	0.21	0.22	0.18	0.24	0.19	0.20	0.16	0.24	0.20
	本文方法	0.72	0.66	0.65	0.58	0.62	0.56	0.63	0.57	0.61	0.51	0.65	0.58
Columbia	MVSS-Net	0.67	0.57	0.63	0.53	0.59	0.49	0.64	0.54	0.37	0.25	0.58	0.48
	PSCC-Net	0.60	0.47	0.59	0.46	0.54	0.42	0.50	0.40	0.17	0.12	0.48	0.37
	TruFor	0.81	0.75	0.77	0.72	0.65	0.56	0.78	0.71	0.45	0.37	0.69	0.62
	MGQFormer	0.78	0.72	0.74	0.69	0.62	0.53	0.75	0.68	0.42	0.34	0.66	0.59
	SAFIRE	0.90	0.84	0.74	0.69	0.72	0.67	0.75	0.70	0.73	0.68	0.77	0.71
	本文方法	0.82	0.76	0.77	0.72	0.75	0.70	0.78	0.72	0.76	0.71	0.78	0.72
Coverage	MVSS-Net	0.45	0.38	0.26	0.18	0.16	0.10	0.13	0.11	0.09	0.06	0.22	0.17
	PSCC-Net	0.45	0.34	0.26	0.19	0.15	0.10	0.14	0.10	0.08	0.05	0.21	0.16
	TruFor	0.56	0.47	0.28	0.23	0.18	0.15	0.30	0.25	0.14	0.10	0.29	0.24
	MGQFormer	0.53	0.45	0.25	0.20	0.15	0.12	0.27	0.22	0.12	0.08	0.26	0.21
	SAFIRE	0.25	0.20	0.22	0.17	0.18	0.14	0.20	0.16	0.16	0.12	0.20	0.16
	本文方法	0.57	0.48	0.55	0.45	0.52	0.42	0.49	0.41	0.40	0.35	0.51	0.42
DSO	MVSS-Net	0.27	0.19	0.24	0.16	0.21	0.15	0.23	0.15	0.20	0.13	0.23	0.15
	PSCC-Net	0.41	0.32	0.37	0.28	0.32	0.22	0.35	0.24	0.29	0.18	0.35	0.25
	TruFor	0.91	0.87	0.80	0.72	0.67	0.55	0.75	0.64	0.45	0.33	0.72	0.62
	MGQFormer	0.88	0.84	0.76	0.68	0.62	0.50	0.71	0.60	0.42	0.30	0.68	0.58
	SAFIRE	0.72	0.68	0.68	0.64	0.63	0.58	0.66	0.62	0.60	0.56	0.66	0.62
	本文方法	0.83	0.73	0.80	0.72	0.75	0.67	0.77	0.70	0.71	0.66	0.77	0.70

Tab. 2 Quantitative comparison of detection results of brightness processed images using F1 and IoU as criteria

数据集	方法	—		$α = 0.2$		$α = 0.5$		$α = - 0.2$		$α = - 0.5$		平均值
数据集	方法	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU
CASIA	MVSS-Net	0.45	0.40	0.40	0.32	0.35	0.29	0.35	0.26	0.25	0.19	0.36	0.29
	PSCC-Net	0.46	0.41	0.42	0.35	0.25	0.18	0.28	0.21	0.22	0.16	0.33	0.26
	TruFor	0.71	0.64	0.47	0.40	0.33	0.27	0.45	0.38	0.35	0.28	0.46	0.39
	MGQFormer	0.68	0.61	0.44	0.37	0.30	0.24	0.42	0.35	0.32	0.25	0.43	0.36
	SAFIRE	0.30	0.24	0.26	0.21	0.22	0.18	0.24	0.19	0.20	0.16	0.24	0.20
	本文方法	0.72	0.66	0.65	0.58	0.62	0.56	0.63	0.57	0.61	0.51	0.65	0.58
Columbia	MVSS-Net	0.67	0.57	0.63	0.53	0.59	0.49	0.64	0.54	0.37	0.25	0.58	0.48
	PSCC-Net	0.60	0.47	0.59	0.46	0.54	0.42	0.50	0.40	0.17	0.12	0.48	0.37
	TruFor	0.81	0.75	0.77	0.72	0.65	0.56	0.78	0.71	0.45	0.37	0.69	0.62
	MGQFormer	0.78	0.72	0.74	0.69	0.62	0.53	0.75	0.68	0.42	0.34	0.66	0.59
	SAFIRE	0.90	0.84	0.74	0.69	0.72	0.67	0.75	0.70	0.73	0.68	0.77	0.71
	本文方法	0.82	0.76	0.77	0.72	0.75	0.70	0.78	0.72	0.76	0.71	0.78	0.72
Coverage	MVSS-Net	0.45	0.38	0.26	0.18	0.16	0.10	0.13	0.11	0.09	0.06	0.22	0.17
	PSCC-Net	0.45	0.34	0.26	0.19	0.15	0.10	0.14	0.10	0.08	0.05	0.21	0.16
	TruFor	0.56	0.47	0.28	0.23	0.18	0.15	0.30	0.25	0.14	0.10	0.29	0.24
	MGQFormer	0.53	0.45	0.25	0.20	0.15	0.12	0.27	0.22	0.12	0.08	0.26	0.21
	SAFIRE	0.25	0.20	0.22	0.17	0.18	0.14	0.20	0.16	0.16	0.12	0.20	0.16
	本文方法	0.57	0.48	0.55	0.45	0.52	0.42	0.49	0.41	0.40	0.35	0.51	0.42
DSO	MVSS-Net	0.27	0.19	0.24	0.16	0.21	0.15	0.23	0.15	0.20	0.13	0.23	0.15
	PSCC-Net	0.41	0.32	0.37	0.28	0.32	0.22	0.35	0.24	0.29	0.18	0.35	0.25
	TruFor	0.91	0.87	0.80	0.72	0.67	0.55	0.75	0.64	0.45	0.33	0.72	0.62
	MGQFormer	0.88	0.84	0.76	0.68	0.62	0.50	0.71	0.60	0.42	0.30	0.68	0.58
	SAFIRE	0.72	0.68	0.68	0.64	0.63	0.58	0.66	0.62	0.60	0.56	0.66	0.62
	本文方法	0.83	0.73	0.80	0.72	0.75	0.67	0.77	0.70	0.71	0.66	0.77	0.70

Tab. 3 Quantitative comparison of detection results of contrast processed images using F1 and IoU as criteria

数据集	方法	—		$β = 0.5$		$β = 0.7$		$β = 1.5$		$β = 2.0$		平均值
数据集	方法	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU
CASIA	MVSS-Net	0.45	0.40	0.34	0.28	0.31	0.25	0.34	0.28	0.32	0.25	0.35	0.29
	PSCC-Net	0.46	0.41	0.33	0.24	0.42	0.32	0.24	0.18	0.33	0.27	0.36	0.28
	TruFor	0.71	0.64	0.45	0.38	0.35	0.28	0.52	0.45	0.46	0.39	0.50	0.43
	MGQFormer	0.68	0.61	0.44	0.37	0.32	0.26	0.50	0.43	0.43	0.36	0.47	0.41
	SAFIRE	0.30	0.24	0.26	0.21	0.23	0.19	0.25	0.20	0.22	0.18	0.25	0.20
	本文方法	0.72	0.66	0.65	0.60	0.62	0.58	0.63	0.58	0.62	0.57	0.65	0.60
Columbia	MVSS-Net	0.67	0.57	0.62	0.50	0.60	0.49	0.62	0.52	0.57	0.47	0.61	0.51
	PSCC-Net	0.60	0.47	0.58	0.46	0.55	0.44	0.57	0.46	0.43	0.33	0.55	0.43
	TruFor	0.81	0.75	0.68	0.60	0.72	0.68	0.77	0.73	0.76	0.71	0.75	0.69
	MGQFormer	0.78	0.72	0.66	0.58	0.70	0.64	0.75	0.71	0.74	0.69	0.73	0.67
	SAFIRE	0.90	0.84	0.74	0.69	0.72	0.67	0.77	0.72	0.75	0.70	0.77	0.72
	本文方法	0.82	0.76	0.76	0.72	0.75	0.69	0.79	0.74	0.77	0.73	0.78	0.73
Coverage	MVSS-Net	0.45	0.38	0.24	0.17	0.30	0.24	0.12	0.08	0.11	0.08	0.24	0.19
	PSCC-Net	0.45	0.34	0.22	0.15	0.29	0.21	0.15	0.11	0.10	0.07	0.24	0.18
	TruFor	0.56	0.47	0.17	0.13	0.24	0.19	0.31	0.26	0.31	0.26	0.32	0.26
	MGQFormer	0.53	0.45	0.15	0.11	0.22	0.17	0.29	0.24	0.28	0.23	0.29	0.24
	SAFIRE	0.25	0.20	0.22	0.17	0.18	0.14	0.20	0.16	0.19	0.15	0.21	0.16
	本文方法	0.57	0.48	0.50	0.39	0.43	0.34	0.53	0.45	0.48	0.39	0.50	0.41
DSO	MVSS-Net	0.27	0.19	0.20	0.13	0.19	0.13	0.23	0.14	0.20	0.14	0.22	0.15
	PSCC-Net	0.41	0.32	0.31	0.20	0.40	0.29	0.34	0.25	0.31	0.22	0.35	0.26
	TruFor	0.91	0.87	0.81	0.74	0.79	0.72	0.77	0.67	0.64	0.51	0.79	0.70
	MGQFormer	0.88	0.84	0.79	0.72	0.76	0.69	0.74	0.65	0.60	0.48	0.75	0.68
	SAFIRE	0.72	0.68	0.70	0.66	0.68	0.64	0.66	0.62	0.62	0.58	0.68	0.64
	本文方法	0.83	0.73	0.81	0.74	0.80	0.74	0.79	0.72	0.75	0.67	0.80	0.72

Tab. 3 Quantitative comparison of detection results of contrast processed images using F1 and IoU as criteria

数据集	方法	—		$β = 0.5$		$β = 0.7$		$β = 1.5$		$β = 2.0$		平均值
数据集	方法	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU
CASIA	MVSS-Net	0.45	0.40	0.34	0.28	0.31	0.25	0.34	0.28	0.32	0.25	0.35	0.29
	PSCC-Net	0.46	0.41	0.33	0.24	0.42	0.32	0.24	0.18	0.33	0.27	0.36	0.28
	TruFor	0.71	0.64	0.45	0.38	0.35	0.28	0.52	0.45	0.46	0.39	0.50	0.43
	MGQFormer	0.68	0.61	0.44	0.37	0.32	0.26	0.50	0.43	0.43	0.36	0.47	0.41
	SAFIRE	0.30	0.24	0.26	0.21	0.23	0.19	0.25	0.20	0.22	0.18	0.25	0.20
	本文方法	0.72	0.66	0.65	0.60	0.62	0.58	0.63	0.58	0.62	0.57	0.65	0.60
Columbia	MVSS-Net	0.67	0.57	0.62	0.50	0.60	0.49	0.62	0.52	0.57	0.47	0.61	0.51
	PSCC-Net	0.60	0.47	0.58	0.46	0.55	0.44	0.57	0.46	0.43	0.33	0.55	0.43
	TruFor	0.81	0.75	0.68	0.60	0.72	0.68	0.77	0.73	0.76	0.71	0.75	0.69
	MGQFormer	0.78	0.72	0.66	0.58	0.70	0.64	0.75	0.71	0.74	0.69	0.73	0.67
	SAFIRE	0.90	0.84	0.74	0.69	0.72	0.67	0.77	0.72	0.75	0.70	0.77	0.72
	本文方法	0.82	0.76	0.76	0.72	0.75	0.69	0.79	0.74	0.77	0.73	0.78	0.73
Coverage	MVSS-Net	0.45	0.38	0.24	0.17	0.30	0.24	0.12	0.08	0.11	0.08	0.24	0.19
	PSCC-Net	0.45	0.34	0.22	0.15	0.29	0.21	0.15	0.11	0.10	0.07	0.24	0.18
	TruFor	0.56	0.47	0.17	0.13	0.24	0.19	0.31	0.26	0.31	0.26	0.32	0.26
	MGQFormer	0.53	0.45	0.15	0.11	0.22	0.17	0.29	0.24	0.28	0.23	0.29	0.24
	SAFIRE	0.25	0.20	0.22	0.17	0.18	0.14	0.20	0.16	0.19	0.15	0.21	0.16
	本文方法	0.57	0.48	0.50	0.39	0.43	0.34	0.53	0.45	0.48	0.39	0.50	0.41
DSO	MVSS-Net	0.27	0.19	0.20	0.13	0.19	0.13	0.23	0.14	0.20	0.14	0.22	0.15
	PSCC-Net	0.41	0.32	0.31	0.20	0.40	0.29	0.34	0.25	0.31	0.22	0.35	0.26
	TruFor	0.91	0.87	0.81	0.74	0.79	0.72	0.77	0.67	0.64	0.51	0.79	0.70
	MGQFormer	0.88	0.84	0.79	0.72	0.76	0.69	0.74	0.65	0.60	0.48	0.75	0.68
	SAFIRE	0.72	0.68	0.70	0.66	0.68	0.64	0.66	0.62	0.62	0.58	0.68	0.64
	本文方法	0.83	0.73	0.81	0.74	0.80	0.74	0.79	0.72	0.75	0.67	0.80	0.72

Fig. 8 Qualitative comparison results

Tab. 4 Ablation study results on different detection modules

检测模块	—		α = 0.5		β = 2.0
检测模块	F1	IoU	F1	IoU	F1	IoU
U-Net	0.432	0.380	0.321	0.298	0.345	0.312
U-Net+DWFEL	0.475	0.428	0.412	0.366	0.409	0.356
U-Net+对比学习	0.652	0.581	0.563	0.505	0.521	0.496
U-Net+DWFEL+对比学习	0.730	0.653	0.659	0.588	0.655	0.582

Tab. 5 Ablation study results on different wavelet bases

小波基	—		α = 0.5		β = 2.0
小波基	F1	IoU	F1	IoU	F1	IoU
Symlet4	0.662	0.694	0.572	0.519	0.545	0.512
Biorthogonal4.4	0.705	0.638	0.632	0.556	0.639	0.560
Haar	0.730	0.653	0.659	0.588	0.655	0.582

Tab. 6 Ablation study results on restoration module

方法	—		α=0.5		β=2.0
方法	F1	IOU	F1	IOU	F1	IOU
无恢复模块	0.691	0.665	0.399	0.345	0.381	0.341
恢复模块	0.730	0.653	0.659	0.588	0.655	0.582

Tab. 7 Ablation study results on different attention mechanisms in restoration module

恢复模块注意力	—		α = 0.5		β = 2.0
恢复模块注意力	F1	IoU	F1	IoU	F1	IoU
Deformable DETR	0.711	0.628	0.631	0.559	0.627	0.554
ViT	0.726	0.649	0.652	0.582	0.653	0.585
Swin Transformer	0.730	0.653	0.659	0.588	0.655	0.582

Tab. 8 Ablation study results on forensics-oriented loss function

方法	—		α = 0.5		β = 2.0
方法	F1	IoU	F1	IoU	F1	IoU
无取证导向损失	0.667	0.611	0.563	0.502	0.542	0.499
取证导向损失	0.730	0.653	0.659	0.588	0.655	0.582

References 37

[1]	FARID H. Image forgery detection［J］. IEEE Signal Processing Magazine， 2009， 26（2）： 16-25.
[2]	STAMM M C， WU M， LIU K J R. Information forensics： an overview of the first decade［J］. IEEE Access， 2013， 1： 167-200.
[3]	BAYAR B， STAMM M C. A deep learning approach to universal image manipulation detection using a new convolutional layer［C］// Proceedings of the 4th ACM Workshop on Information Hiding and Multimedia Security. New York： ACM， 2016： 5-10.
[4]	王宏，钱清，王欢，等. 融合大核注意力卷积的轻量化图像篡改定位算法［J］. 计算机应用， 2023， 43（9）： 2692-2699.
	WANG H， QIAN Q， WANG H， et al. Lightweight image tamper localization algorithm based on large kernel attention convolution［J］. Journal of Computer Applications， 2023， 43（9）： 2692-2699.
[5]	XU Y， REN J， FANG A， et al. AFCMS-Net： adaptive feature coupling and multi-level supervision network for effective image forgery localization［J］. Knowledge-Based Systems， 2025， 327： No.114126.
[6]	VIJAYALAKSHMI D， NATH M K， ACHARYA O P. A comprehensive survey on image contrast enhancement techniques in spatial domain［J］. Sensing and Imaging， 2020， 21： No.40.
[7]	COZZOLINO D， VERDOLIVA L. Noiseprint： a CNN-based camera model fingerprint［J］. IEEE Transactions on Information Forensics and Security， 2020， 15： 144-159.
[8]	LIU Y， CHEN S， SHI H， et al. MUN： image forgery localization based on M³ encoder and UN decoder［C］// Proceedings of the 39th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2025： 5685-5693.
[9]	TREBING K， STAŃCZYK T， MEHRKANOON S. SmaAt-UNet： precipitation nowcasting using a small attention-UNet architecture［J］. Pattern Recognition Letters， 2021， 145： 178-186.
[10]	DONG J， WANG W， TAN T. CASIA image tampering detection evaluation database［C］// Proceedings of the 2013 IEEE China Summit and International Conference on Signal and Information Processing. Piscataway： IEEE， 2013： 422-426.
[11]	HSU Y F， CHANG S F. Detecting image splicing using geometry invariants and camera characteristics consistency［C］// Proceedings of the 2006 IEEE International Conference on Multimedia and Expo. Piscataway： IEEE， 2006： 549-552.
[12]	WEN B， ZHU Y， SUBRAMANIAN R， et al. COVERAGE： a novel database for copy-move forgery detection ［C］// Proceedings of the 2016 IEEE International Conference on Image Processing. Piscataway： IEEE， 2016： 161-165.
[13]	DE CARVALHO T J， RIESS C， ANGELOPOULOU E， et al. Exposing digital image forgeries by illumination color classification［J］. IEEE Transactions on Information Forensics and Security， 2013， 8（7）： 1182-1194.
[14]	SU J， XU B， YIN H. A survey of deep learning approaches to image restoration［J］. Neurocomputing， 2022， 487： 46-65.
[15]	ZAMIR S W， ARORA A， KHAN S， et al. Multi-stage progressive image restoration［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 14816-14826.
[16]	SAHARIA C， HO J， CHAN W， et al. Image super-resolution via iterative refinement［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（4）： 4713-4726.
[17]	ZANARDELLI M， GUERRINI F， LEONARDI R， et al. Image forgery detection： a survey of recent deep-learning approaches［J］. Multimedia Tools and Applications， 2023， 82（12）： 17521-17566.
[18]	ZHU J， LI D， FU X， et al. Learning discriminative noise guidance for image forgery detection and localization［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2024： 7739-7747.
[19]	LI S， MA W， GUO J， et al. UnionFormer： unified-learning Transformer with multi-view representation for image manipulation detection and localization［C］// Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2024： 12523-12533.
[20]	RAO Y， NI J. Self-supervised domain adaptation for forgery localization of JPEG compressed images［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 15014-15023.
[21]	ZHUO L， TAN S， LI B， et al. Self-adversarial training incorporating forgery attention for image forgery localization［J］. IEEE Transactions on Information Forensics and Security， 2022， 17： 819-834.
[22]	WU H， ZHOU J， TIAN J， et al. Robust image forgery detection over online social network shared images［C］// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2022： 13430-13439.
[23]	HOWARD A， SANDLER M， CHU G， et al. Searching for MobileNetV3［C］// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2019： 1314-1324.
[24]	TAN M， LE Q V. EfficientNetv2： smaller models and faster training［C］// Proceedings of the 38th International Conference on Machine Learning. New York： JMLR.org， 2021： 10096-10106.
[25]	CAI Y， BIAN H， LIN J， et al. Retinexformer： one-stage Retinex-based Transformer for low-light image enhancement［C］// Proceedings of the 2023 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2023： 12470-12479.
[26]	AFIFI M， DERPANIS K G， OMMER B， et al. Learning multi-scale photo exposure correction［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 9153-9163.
[27]	DAI J， QI H， XIONG Y， et al. Deformable convolutional networks［C］// Proceedings of the 2017 IEEE International Conference on Computer Vision. Piscataway： IEEE， 2017： 764-773.
[28]	ZHAO X， VEMULAPALLI R， MANSFIELD P A， et al. Contrastive learning for label efficient semantic segmentation［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 10603-10613.
[29]	JIANG L， DAI B， WU W， et al. Focal frequency loss for image reconstruction and synthesis［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 13899-13909.
[30]	MUSTAFA A， MIKHAILIUK A， ILIESCU D A， et al. Training a task-specific image reconstruction loss［C］// Proceedings of the 2022 IEEE/CVF Winter Conference on Applications of Computer Vision. Piscataway： IEEE， 2022： 21-30.
[31]	RONG W， LI Z， ZHANG W， et al. An improved Canny edge detection algorithm［C］// Proceedings of the 2014 IEEE International Conference on Mechatronics and Automation. Piscataway： IEEE， 2014： 577-582.
[32]	GUAN H， KOZAK M， ROBERTSON E， et al. MFC datasets： large-scale benchmark datasets for media forensic challenge evaluation［C］// Proceedings of the 2019 IEEE Winter Applications of Computer Vision Workshops. Piscataway： IEEE， 2019： 63-72.
[33]	DONG C， CHEN X， HU R， et al. MVSS-Net： multi-view multi-scale supervised networks for image manipulation detection［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（3）： 3539-3553.
[34]	LIU X， LIU Y， CHEN J， et al. PSCC-Net： progressive spatio-channel correlation network for image manipulation detection and localization［J］. IEEE Transactions on Circuits and Systems for Video Technology， 2022， 32（11）： 7505-7517.
[35]	GUILLARO F， COZZOLINO D， SUD A， et al. TruFor： leveraging all-round clues for trustworthy image forgery detection and localization［C］// Proceedings of the 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2023： 20606-20615.
[36]	ZENG K， CHENG R， TAN W， et al. MGQFormer： mask-guided query-based Transformer for image manipulation localization［C］// Proceedings of the 38th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2024： 6944-6952.
[37]	KWON M J， LEE W， NAM S H， et al. SAFIRE： segment any forged image region［C］// Proceedings of the 39th AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2025： 4437-4445.

Image tampering localization and detection network under brightness-contrast disturbances

亮度对比度扰动下的图像篡改定位检测网络

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 16

References 37

Related Articles 15

Recommended Articles

Metrics

[1]	Jinghong WANG, Xiao CHEN, Yingmei MA, Bi LI, Jusheng MI, Wei WANG. Multi-level neighborhood contrastive attribute graph clustering based on adaptive learning [J]. Journal of Computer Applications, 2026, 46(6): 1836-1843.
[2]	Bingqing LI, Binhao HUANG, Yubei TANG, Baili ZHANG. Quality of service prediction model for data sparsity and cold start problems [J]. Journal of Computer Applications, 2026, 46(6): 1829-1835.
[3]	Hang QI, Tingting DONG, Yongqiang NAI, Xian MO. Contrastive collaborative filtering method based on graph diffusion generation and adaptive sampling [J]. Journal of Computer Applications, 2026, 46(6): 1818-1828.
[4]	Kun FU, Haoyu WEI, Weijing LIU, Xing DANG, Zezheng LIU, Jianwei LI. Graph neural network framework for topology semantic dual-domain collaboration [J]. Journal of Computer Applications, 2026, 46(5): 1378-1387.
[5]	Dirui ZHANG, Jiayu LIN, Zuhong LIANG. Supervised contrastive generative sentiment analysis method with uncertainty-aware unlikelihood learning [J]. Journal of Computer Applications, 2026, 46(5): 1416-1423.
[6]	Haihua ZHAO, Yijun HU, Rui TANG, Xian MO. Multimodal recommendation method based on semantic fusion and contrast enhancement [J]. Journal of Computer Applications, 2026, 46(4): 1058-1068.
[7]	Yuhang XIAO, Guanfeng LI, Yuyin CHEN, Jing QIN. Few-shot relation extraction model with graph-based multi-view contrastive learning [J]. Journal of Computer Applications, 2026, 46(3): 732-740.
[8]	Xiaoxia LIU, Liqun KUANG, Song WANG, Shichao JIAO, Huiyan HAN, Fengguang XIONG. Multi-scale spatio-temporal decoupling for contrastive learning of skeleton action recognition [J]. Journal of Computer Applications, 2026, 46(3): 767-774.
[9]	Limei DONG, Yanzi LI, Jiayin LI, Li XU. Neighborhood-enhanced unsupervised graph anomaly detection [J]. Journal of Computer Applications, 2026, 46(2): 458-466.
[10]	Hu LUO, Mingshu ZHANG. Rumor detection method based on cross-modal attention mechanism and contrastive learning [J]. Journal of Computer Applications, 2026, 46(2): 361-367.
[11]	Ziyang CHENG, Ruizhang HUANG, Jingjing XUE. Deep evolutionary topic clustering model [J]. Journal of Computer Applications, 2026, 46(1): 85-94.
[12]	Wen LI, Kairong LI, Kai YANG. Subgraph-aware contrastive learning with data augmentation [J]. Journal of Computer Applications, 2026, 46(1): 1-9.
[13]	Xingyao YANG, Zheng QI, Jiong YU, Zulian ZHANG, Shuai MA, Hongtao SHEN. Session-based recommendation model based on time-aware and space-enhanced dual channel graph neural network [J]. Journal of Computer Applications, 2026, 46(1): 104-112.
[14]	Zhixiong XU, Bo LI, Xiaoyong BIAN, Qiren HU. Adversarial sample embedded attention U-Net for 3D medical image segmentation [J]. Journal of Computer Applications, 2025, 45(9): 3011-3016.
[15]	Chao LIU, Yanhua YU. Knowledge-aware recommendation model combining denoising strategy and multi-view contrastive learning [J]. Journal of Computer Applications, 2025, 45(9): 2827-2837.