A recurrent dynamic model for correspondence-based face recognition

Abstract

Our aim here is to create a fully neural, functionally competitive, and correspondence-based model for invariant face recognition. By recurrently integrating information about feature similarities, spatial feature relations, and facial structure stored in memory, the system evaluates face identity ("what"-information) and face position ("where"-information) using explicit representations for both. The network consists of three functional layers of processing, (1) an input layer for image representation, (2) a middle layer for recurrent information integration, and (3) a gallery layer for memory storage. Each layer consists of cortical columns as functional building blocks that are modeled in accordance with recent experimental findings. In numerical simulations we apply the system to standard benchmark databases for face recognition. We find that recognition rates of our biologically inspired approach lie in the same range as recognition rates of recent and purely functionally motivated systems.

Dynamic Coordination in Brain and Mind

“…Most importantly, however, it is inherently a relational concept as it is concerned with connections between cells, rather than simply with the activity of the cells themselves. This therefore resonates with the idea of the rapid formation, matching, and dissolution of effective network architectures as used within dynamic link architectures (Wolfrum et al 2008).…”

“…Such recognition amounts to graph matching, i.e., to the comparison of a stored model of the pattern and the potential instance to be recognized, where both have the form of a graph with feature-labelled nodes and links to represent neighbourhood relationships. During the graph matching, or recognition, process dynamic links between corresponding nodes have to be established, as modelled in Wolfrum et al (2008).…”

Information Routing, Correspondence Finding, and Object Recognition in the Brain

“…We report the behavior of the network and test it for different tasks in Section 2.5, before concluding the chapter in Section 2.6. The contents of this chapter were partially published in von der Malsburg 2008, Wolfrum, Wolff, Lücke and, the material presented in Section 2.5.4 in .…”

A correspondence-based neural mechanism for position invariant feature processing

“…Then, taking a maximum operation of the similarity measures, the most appropriate projections are detected to establish a whole connection between the input and model layers, specifying the object position on the input image. Finally, employing a dynamic model of cortical columns [3], we propose a position-invariant object recognition system in a dynamic routing circuit, without any loss of concepts about the position-specific marginalized features mentioned above. Then, we will test and discuss the ability of our proposed system for recognition performance, specifying a correct position of a particular object.…”