Seal Extraction and Incomplete Restoration in Complex Document Image
Diheng Sun · 2021
This paper proposes a method to quickly locate and extract the seal from the complex background. First, this paper uses the color model and two-dimensional projection method to quickly distinguish the local block(s) with red seals from other red interference information. This step not only reduces the complexity of stamp extraction work, but also increases the accuracy of positioning. Then use the specific shape feature of the seal to locate the seal image. Then, the difference in color is used to separate the seal from the background covered by the seal. Finally, use methods such as filling to repair the incomplete seal. Experimental results show that the proposed method can better locate, extract and repair the seal in most cases, and complete the entire task faster than most of the compared techniques.