<s>
The	O
Hessian	B-Algorithm
affine	I-Algorithm
region	I-Algorithm
detector	I-Algorithm
is	O
a	O
feature	O
detector	O
used	O
in	O
the	O
fields	O
of	O
computer	B-Application
vision	I-Application
and	O
image	B-General_Concept
analysis	I-General_Concept
.	O
</s>
<s>
Like	O
other	O
feature	O
detectors	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
is	O
typically	O
used	O
as	O
a	O
preprocessing	O
step	O
to	O
algorithms	O
that	O
rely	O
on	O
identifiable	O
,	O
characteristic	O
interest	O
points	O
.	O
</s>
<s>
The	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
is	O
part	O
of	O
the	O
subclass	O
of	O
feature	O
detectors	O
known	O
as	O
affine-invariant	O
detectors	O
:	O
Harris	B-Algorithm
affine	I-Algorithm
region	I-Algorithm
detector	I-Algorithm
,	O
Hessian	B-Algorithm
affine	I-Algorithm
regions	O
,	O
maximally	B-Algorithm
stable	I-Algorithm
extremal	I-Algorithm
regions	I-Algorithm
,	O
Kadir	B-Algorithm
–	I-Algorithm
Brady	I-Algorithm
saliency	I-Algorithm
detector	I-Algorithm
,	O
edge-based	O
regions	O
(	O
EBR	O
)	O
and	O
intensity-extrema-based	O
(	O
IBR	O
)	O
regions	O
.	O
</s>
<s>
The	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
algorithm	O
is	O
almost	O
identical	O
to	O
the	O
Harris	B-Algorithm
affine	I-Algorithm
region	I-Algorithm
detector	I-Algorithm
.	O
</s>
<s>
The	O
Harris	O
affine	O
detector	O
relies	O
on	O
interest	O
points	O
detected	O
at	O
multiple	O
scales	O
using	O
the	O
Harris	B-Algorithm
corner	I-Algorithm
measure	I-Algorithm
on	O
the	O
second-moment	O
matrix	O
.	O
</s>
<s>
The	O
Hessian	B-Algorithm
affine	I-Algorithm
also	O
uses	O
a	O
multiple	O
scale	O
iterative	O
algorithm	O
to	O
spatially	O
localize	O
and	O
select	O
scale	O
and	O
affine	O
invariant	O
points	O
.	O
</s>
<s>
However	O
,	O
at	O
each	O
individual	O
scale	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
chooses	O
interest	O
points	O
based	O
on	O
the	O
Hessian	O
matrix	O
at	O
that	O
point	O
:	O
</s>
<s>
As	O
discussed	O
in	O
the	O
Harris	B-Algorithm
affine	I-Algorithm
region	I-Algorithm
detector	I-Algorithm
article	O
,	O
the	O
derivatives	O
must	O
be	O
scaled	O
appropriately	O
by	O
a	O
factor	O
related	O
to	O
the	O
Gaussian	O
kernel	O
:	O
.	O
</s>
<s>
This	O
type	O
of	O
measure	O
is	O
very	O
similar	O
to	O
the	O
measures	O
used	O
in	O
the	O
blob	B-Algorithm
detection	I-Algorithm
schemes	O
proposed	O
by	O
Lindeberg	O
(	O
1998	O
)	O
,	O
where	O
either	O
the	O
Laplacian	O
or	O
the	O
determinant	O
of	O
the	O
Hessian	O
were	O
used	O
in	O
blob	B-Algorithm
detection	I-Algorithm
methods	O
with	O
automatic	O
scale	O
selection	O
.	O
</s>
<s>
Furthermore	O
,	O
using	O
these	O
initially	O
detected	O
points	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
uses	O
an	O
iterative	O
shape	O
adaptation	O
algorithm	O
to	O
compute	O
the	O
local	O
affine	O
transformation	O
for	O
each	O
interest	O
point	O
.	O
</s>
<s>
The	O
implementation	O
of	O
this	O
algorithm	O
is	O
almost	O
identical	O
to	O
that	O
of	O
the	O
Harris	O
affine	O
detector	O
;	O
however	O
,	O
the	O
above	O
mentioned	O
Hessian	O
measure	O
replaces	O
all	O
instances	O
of	O
the	O
Harris	B-Algorithm
corner	I-Algorithm
measure	I-Algorithm
.	O
</s>
<s>
(	O
2005	O
)	O
have	O
done	O
a	O
thorough	O
analysis	O
of	O
several	O
state	O
of	O
the	O
art	O
affine	O
region	O
detectors	O
:	O
Harris	O
affine	O
,	O
Hessian	B-Algorithm
affine	I-Algorithm
,	O
MSER	B-Algorithm
,	O
IBR	O
&	O
EBR	O
and	O
salient	B-Algorithm
detectors	O
.	O
</s>
<s>
Overall	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
performs	O
second	O
best	O
to	O
MSER	B-Algorithm
.	O
</s>
<s>
Like	O
the	O
Harris	O
affine	O
detector	O
,	O
Hessian	B-Algorithm
affine	I-Algorithm
interest	O
regions	O
tend	O
to	O
be	O
more	O
numerous	O
and	O
smaller	O
than	O
other	O
detectors	O
.	O
</s>
<s>
For	O
a	O
single	O
image	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
typically	O
identifies	O
more	O
reliable	O
regions	O
than	O
the	O
Harris-Affine	B-Algorithm
detector	O
.	O
</s>
<s>
The	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
responds	O
well	O
to	O
textured	O
scenes	O
in	O
which	O
there	O
are	O
a	O
lot	O
of	O
corner-like	O
parts	O
.	O
</s>
<s>
However	O
,	O
for	O
some	O
structured	O
scenes	O
,	O
like	O
buildings	O
,	O
the	O
Hessian	B-Algorithm
affine	I-Algorithm
detector	O
performs	O
very	O
well	O
.	O
</s>
<s>
This	O
is	O
complementary	O
to	O
MSER	B-Algorithm
that	O
tends	O
to	O
do	O
better	O
with	O
well	O
structured	O
(	O
segmentable	O
)	O
scenes	O
.	O
</s>
<s>
:	O
K	O
.	O
Mikolajczyk	O
maintains	O
a	O
web	O
page	O
that	O
contains	O
Linux	O
binaries	O
of	O
the	O
Hessian-Affine	B-Algorithm
detector	O
in	O
addition	O
to	O
other	O
detectors	O
and	O
descriptors	O
.	O
</s>
