Real-time Quadrilateral Object Corner Detection Algorithm Based on Deep Learning

In the field of computer vision, there are many quadrilateral objects such as slide, screen, book, etc, which usually has the characteristics of rotation, perspective, and beveling. Therefore, it becomes a key task to detect the corner points of the quadrilateral object to restore the content. This paper proposes an algorithm to solve the problem of quadrilateral object corner detection. Different from the previous methods, our method does not make any assumptions about the background, distance, content and other attributes of the quadrilateral object, which can be applied to a wider range of quadrilateral objects. We named the proposed method as Corner Detect Network (CDN). Simulation results show that the proposed algorithm can quickly and accurately extract the corner point of quadrilateral object.