Context. Thanks to their extensive and homogeneous sky coverage, deep, large-scale, multiwavelength surveys are uniquely suited to statistically identify and map young star clusters in our Galaxy. Such studies are crucial to address issues like the initial mass function, or the modes and dynamics of star cluster formation and evolution. Aims. We aim to test a purely photometric approach to statistically identify a young clustered population embedded in a large population of field stars, with no prior knowledge of the nature of stars in the field. We conducted our blind test study on the NGC 2264 region, which hosts a well-known, richly populated young cluster (∼3 Myr-old) and several active star-forming sites. Methods. We selected a large (4 deg²) area around the NGC 2264 cluster, and assembled an extensive r, i, J catalog of the field from pre-existing large-scale surveys, notably Pan-STARRS1 and UKIDSS. We then mapped the stellar color locus on the (i – J, r – i) diagram to select M-type stars, which offer the following observational advantages with respect to more massive stars: (i) they comprise a significant fraction of the Galactic stellar population; (ii) their pre-main sequence phase lasts significantly longer than for higher mass stars; (iii) they exhibit the strongest luminosity evolution from the pre-main sequence to the main sequence; (iv) their observed r, i, J colors provide a direct and empirical estimate of A_V. A comparative analysis of the photometric and spatial properties of M-type stars as a function of A_V enabled us to probe the structure and stellar content of our field. Results. Using only r, i, J photometry, we could identify two distinct populations in our field: a diffuse field population and a clustered population in the center of the field. The presence of a concentration of occulting material, spatially associated with the clustered population, allowed us to derive an estimate of its distance (800–900 pc) and age (∼0.5–5 Myr); these values are overall consistent with the literature parameters for the NGC 2264 star-forming region. The extracted clustered population exhibits a hierarchical structure, with two main clumps and peaks in number density of objects around the most reddened locations within the field. An excellent agreement is found between the observed substructures for the clustered population and a map of the NGC 2264 subregions reported in the literature. Our selection of clustered members is coherent with the literature census of the NGC 2264 cluster for about 95% of the objects located in the inner regions of the field, where the estimated contamination rate by field stars in our sample is only 2%. In addition, the availability of a uniform dataset for a large area around the NGC 2264 region enabled us to discover a population of about a hundred stars with indications of statistical membership to the cluster, therefore extending the low-mass population census of NGC 2264 to distances of 10–15 pc from the cluster cores. Conclusions. By making use solely of deep, multiband (r, i, J) photometry, without assuming any further knowledge of the stellar population of our field, we were able to statistically identify and reconstruct the structure of a very young cluster that has been a prime target for star formation studies over several decades. The method tested here can be readily applied to surveys such as Pan-STARRS and the future LSST to undertake a first complete census of low-mass, young stellar populations down to distances of several kiloparsecs across the Galactic plane.