Content that shows up in pictures contains significant and helpful in-arrangement. Location and extraction of content in pictures have been utilized in numerous applications. In this paper, we propose a multi-scale edge-based content extraction calculation, which can naturally distinguish and remove message in complex pictures. The proposed technique is a broadly useful content location and ex-footing calculation, which can bargain with printed record pictures as well as with scene content. It is powerful concerning the text dimension, style, shading, direction, and arrangement of content and can be utilized in a huge assortment of utilization fields, for example, versatile robot route, vehicle permit discovery and acknowledgment, object identification, archive recovering, page division, and so forth.