arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.04173cs.CV
收藏

利用Segment Anything进行边缘信息学习的机器图像编码

Image Coding for Machines with Edge Information Learning Using Segment Anything

Graduate School of Fundamental Science and Engineering, Waseda University(早稻田大学基础科学与工程研究院);Waseda University(早稻田大学)

Takahiro Shindo, Kein Yamada, Taiju Watanabe, Hiroshi Watanabe

AI总结:

本文提出SA-ICM,利用Segment Anything生成的边缘信息训练图像压缩模型,仅编码物体边缘,在提升图像识别压缩性能的同时保护隐私,并扩展至视频压缩模型SA-NeRV。

摘要:

面向机器的图像编码(Image Coding for Machines,ICM)是一种用于图像识别的图像压缩技术。由于对图像识别人工智能的需求日益增长,这项技术变得至关重要。在本文中,我们提出了一种ICM方法,该方法专注于仅对图像中物体部分的边缘信息进行编码和解码,我们称之为SA-ICM。这是一个使用由Segment Anything生成的边缘信息训练的学习型图像压缩(Learned Image Compression,LIC)模型。我们的方法可用于具有各种任务的图像识别模型。SA-ICM对输入数据的变化也具有鲁棒性,使其在多种使用场景中有效。此外,我们的方法从隐私角度提供了益处,因为它在编码器端移除了人脸信息,从而保护了个人隐私。此外,这种LIC模型训练方法可用于训练神经视频表示(Neural Representations for Videos,NeRV),这是一种视频压缩模型。通过使用由Segment Anything生成的边缘信息训练NeRV,可以创建对图像识别有效的NeRV(SA-NeRV)。实验结果证实了SA-ICM的优势,在面向图像识别的图像压缩中呈现出最佳性能。我们还表明,SA-NeRV在面向机器的视频压缩中优于普通NeRV。代码可在https://github.com/final-0/SA-ICM获取。

英文摘要:

Image Coding for Machines (ICM) is an image compression technique for image recognition. This technique is essential due to the growing demand for image recognition AI. In this paper, we propose a method for ICM that focuses on encoding and decoding only the edge information of object parts in an image, which we call SA-ICM. This is an Learned Image Compression (LIC) model trained using edge information created by Segment Anything. Our method can be used for image recognition models with various tasks. SA-ICM is also robust to changes in input data, making it effective for a variety of use cases. Additionally, our method provides benefits from a privacy point of view, as it removes human facial information on the encoder's side, thus protecting one's privacy. Furthermore, this LIC model training method can be used to train Neural Representations for Videos (NeRV), which is a video compression model. By training NeRV using edge information created by Segment Anything, it is possible to create a NeRV that is effective for image recognition (SA-NeRV). Experimental results confirm the advantages of SA-ICM, presenting the best performance in image compression for image recognition. We also show that SA-NeRV is superior to ordinary NeRV in video compression for machines. Code is available at https://github.com/final-0/SA-ICM.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑