Multi-resolution network based image steganalysis model

What is it about?

this paper proposes a multi-resolution steganalysis net (MRS-Net). MRS-Net adopts a multi-resolution network to extract global image information, fusing the output feature map to ensure high-dimensional semantic information and supplementing low-level detail information. Furthermore, the model incorporates an attention module which can analyze image sensitivity based on different channel and spatial information, thus effectively focusing on areas with rich steganographic signals

Featured Image

Why is it important?

this paper proposes a multi-resolution steganalysis net (MRS-Net). MRS-Net adopts a multi-resolution network to extract global image information, fusing the output feature map to ensure high-dimensional semantic information and supplementing low-level detail information. Furthermore, the model incorporates an attention module which can analyze image sensitivity based on different channel and spatial information, thus effectively focusing on areas with rich steganographic signals

Perspectives