AbstractFor the Hilbert type multiple integral inequality $$ \int _{\mathbb{R}_{+}^{n}} \int _{\mathbb{R}_{+}^{m}} K\bigl( \Vert x \Vert _{m,\rho }, \Vert y \Vert _{n, \rho }\bigr) f(x)g(y) \,\mathrm{d} x \,\mathrm{d} y \leq M \Vert f \Vert _{p, \alpha } \Vert g \Vert _{q, \beta } $$
∫
R
+
n
∫
R
+
m
K
(
∥
x
∥
m
,
ρ
,
∥
y
∥
n
,
ρ
)
f
(
x
)
g
(
y
)
d
x
d
y
≤
M
∥
f
∥
p
,
α
∥
g
∥
q
,
β
with a nonhomogeneous kernel $K(\|x\|_{m, \rho }, \|y\|_{n, \rho })=G(\|x\|^{\lambda _{1}}_{m, \rho }/ \|y\|^{\lambda _{2}}_{n, \rho })$
K
(
∥
x
∥
m
,
ρ
,
∥
y
∥
n
,
ρ
)
=
G
(
∥
x
∥
m
,
ρ
λ
1
/
∥
y
∥
n
,
ρ
λ
2
)
($\lambda _{1}\lambda _{2}> 0$
λ
1
λ
2
>
0
), in this paper, by using the weight function method, necessary and sufficient conditions that parameters p, q, $\lambda _{1}$
λ
1
, $\lambda _{2}$
λ
2
, α, β, m, and n should satisfy to make the inequality hold for some constant M are established, and the expression formula of the best constant factor is also obtained. Finally, their applications in operator boundedness and operator norm are also considered, and the norms of several integral operators are discussed.
Multiscale Detection of Circles, Ellipses and Line Segments, Robust to Noise and Blur