51ÁÔÆæÈë¿Ú

Abstract:

We report on the development of a new method for detecting automatically malignant masses in digitised mammograms, based on a novel non-linear method of multiscale analysis. We show that it is possible to detect > 85 percent of all malignant masses by this method, irrespective of their size, with a false-alarm rate that is close to acceptable for asymptomatic screening.

Abstract:

We present a redshift survey of 124 galaxies, from an imaging $K$-band survey complete to $K\simeq 17.3$. The optical-to-infrared colours are consistent with the range expected from synthetic galaxy spectra, although there are some cases of very red nuclei. Our data show no evidence for evolution of the $K$-band luminosity function at $z<0.5$, and the results are well described by a Schechter function with $M_K^*=-22.75\pm0.13+5\log_{10}h$ and $\phi^*=0.026\pm0.003 h^3 {\rm Mpc^{-3}}$. This is a somewhat higher normalization than has been found by previous workers, and it removes much of the excess in faint $K$ and $B$ counts with respect to a no-evolution model. However, we do find evidence for evolution at $z>0.5$: $M_K^*$ is approximately 0.75 mag. brighter at $z=1$. This luminosity evolution is balanced by a reduced normalization at high redshift. The overall evolution is thus opposite to that expected in simple merger-dominated models.