高级检索

基于深度鉴伪的图像和谐化方法

Image Harmonization Method Based on Deep Forgery Detection

  • 摘要: 在计算机视觉和增强现实领域中, 将前景物体融合到背景场景中并实现图像和谐是一项重要且具有挑战性的任务. 目前和谐化方法大多通过调整图像前景的外观使其与背景相适应来达到视觉上一致性, 和谐化效果的提升空间有限. 为了进一步提升图像和谐化性能, 所提方法通过引入深度鉴伪网络, 利用生成对抗机制将鉴伪网络的结果作为判定指标与现有编码器-解码器和谐化网络构建 GAN 模型, 二者互相博弈以达到鉴伪网络无法识别出和谐重构后的结果为合成图像的目的, 并在鉴伪网络中添加了图像差分模块和图像光照模块, 该部分有效地增强其鉴定性能. 从定性和定量 2 方面在公共的基准数据集 iHarmony4 上进行多组实验, 结果表明, 所提方法均优于对比方法, 其中, 比性能最好的对比方法在 MSE 和 PSNR 评估指标上分别提高了 0.24 和 0.16 dB, 在图像和谐化任务上取得了优异的表现.

     

    Abstract: In computer vision and augmented reality fields, it is an important and challenging task to fuse foreground objects into the background scene and achieve image harmonization. Most of the current mainstream harmonization methods adjust the appearance of the image foreground to make it compatible with the background visually, but the room for improving the harmonization effect is limited. In order to further improve the performance, this paper proposes an image harmonization authenticity identification network. Construct a GAN model with the results of the forgery detection network as the judgment indicator and the existing encoder-decoder harmonization network using the generative adversarial mechanism. The two networks compete with each other to achieve the result that the authenticity identification network cannot recognize the reconstruction result of the harmonization network as a synthetic image. The image difference module and image illumination module are added to the counterfeiting identification network, which effectively enhances its identification performance. Qualitative and quantitative experiments were conducted on the public benchmark dataset iHarmony4. The results demonstrate that the proposed method outperforms all comparative methods. In particular, compared to the best-performing comparative method, there were improvements of 0.24 and 0.16dB in the MSE and PSNR, and this method has achieved excellent performance in image harmonization tasks.

     

/

返回文章
返回