Los criterios diagnosticos de IDCV fueron establecidos por las sociedades Europea y Panamericana de inmunodeficiencias (ESID/ PAGID) en 1999 y siguen vigentes hasta el momento, incluyen: Disminucion marcada (al menos 2 desviaciones estandar por debajo del promedio para la edad) de IgG, IgA y/o IgM en suero; ser-mayor de cuatro anos, ser negativo para isohemaglutininas y/o tener una pobre respuesta a vacunas, ademas que se hayan excluido otras causas de hipogammaglobulinemia.
Es importante resaltar que hasta el 25% de los pacientes con IDCV cursan con conteos de LT [CD4.sup.+] bajos, sin tener un compromiso de la inmunidad celular.
ND is a newly added feature of IDCV algorithm, namely, the centroid distance of the biggest connected coherence pixels.
The similarity of two images for IDCV algorithm computed from (5) is [D.sub.IDCV] = 442.6, where the value of [omega] is 0.7.
It can be seen from [D.sub.DCV] = 558 and [D.sub.IDCV] = 442.6 that the difference of similar images polluted by noise adopting the DCV or IDCV algorithm is comparatively larger; thus DCV and IDCV algorithms are with poorer performance of antinoise.
As the image retrieval algorithm based on DCV or IDCV is relatively more sensitive to noise, a higher false retrieval rate will be resulted in by image retrieval for the image database seriously polluted by noise.
It illustrates that the DCV and IDCV algorithms are sensitive to noise in Section 2.2 and different shapes may have the same descriptor.
The similarity of Figures 7(a) and 7(b) for IDCV algorithm is computed from (3) as follows:
To illustrate the superiority of the algorithm, a comparison is made among the MDCV algorithm and DCV and IDCV algorithms.
Figure 11 shows us the contrast curves for the precision rate and recall rate for IDCV, DCV, and MDCV algorithms.
In the experiment, by assuming that the scale factors [[sigma].sub.1] = 5 and [[sigma].sub.2] = 20 and [[omega].sub.1] = 0.2 and [[omega].sub.2] = 0.7, here, we still compare MDCV with DCV and IDCV algorithms.
It can be seen from the contrast curves of this experiment that MDCV algorithm is obviously superior to DCV and IDCV algorithms for retrieval of the image database which is badly polluted by noise.