Visual place recognition (VPR) is the problem of recognising a previously\nvisited location using visual information. Many attempts to improve the\nperformance of VPR methods have been made in the literature. One approach that\nhas received attention recently is the multi-process fusion where different VPR\nmethods run in parallel and their outputs are combined in an effort to achieve\nbetter performance. The multi-process fusion, however, does not have a\nwell-defined criterion for selecting and combining different VPR methods from a\nwide range of available options. To the best of our knowledge, this paper\ninvestigates the complementarity of state-of-the-art VPR methods systematically\nfor the first time and identifies those combinations which can result in better\nperformance. The paper presents a well-defined framework which acts as a sanity\ncheck to find the complementarity between two techniques by utilising a\nMcNemar's test-like approach. The framework allows estimation of upper and\nlower complementarity bounds for the VPR techniques to be combined, along with\nan estimate of maximum VPR performance that may be achieved. Based on this\nframework, results are presented for eight state-of-the-art VPR methods on ten\nwidely-used VPR datasets showing the potential of different combinations of\ntechniques for achieving better performance.\n