outlines randomly modified #235
Comments
|
Hi there,
I don't think that's a matter of shape complexity but how the masks were
done. The current algorithm is only fully reproducible when the masks are
on the middle point of the jpg. If not, the algo will try to find a black
pixel, randomly, on the jpg. Also when single pixels are isolated on the
diagonal eg the top right 'x' below:
```
x o x
x x o
x o o
```
outline extract may/will fail
hope this helps,
if not, send me the masks that produce different outlines and I gonna have
a look,
all the best,
v
Le lun. 24 juin 2024 à 10:59, raynernunez ***@***.***> a
écrit :
… Hi, I'm trying to compare these outlines, Lepidoptera genitalia, but for
some reason momocs gets wrong, and randomly, some of the images. I removed
part of the structure in some samples to compare and the outlines were
captured fine, look at the two marked samples. Maybe these shapes are too
complex? Any help would be great. Thanks.
body.only.jpg (view on web)
<https://github.com/MomX/Momocs/assets/97603334/4b832afb-f285-4f95-8d5a-a32cda647f05>
complete.jpg (view on web)
<https://github.com/MomX/Momocs/assets/97603334/3180c57e-fa30-4361-b741-f319a4fb6b1e>
—
Reply to this email directly, view it on GitHub
<#235>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABECQFQBOQES4IJODZZFOG3ZI7NX3AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM3DSNRVGMZDIMQ>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
--
vincentbonhomme.fr <http://www.vincentbonhomme.fr>
cevennette.fr <https://www.cevennette.fr> 🌱
|
|
I'll check that, thank you very much |
|
Hi, when I created the mask I cropped them at the same size and tried them
to be at the center. Now I tried adding a bit more white background around
the shapes but it is still happening and each time I run it gets erroneous
outlines for different samples (attached 2 examples).
Here are the masks so you can try to figure out how to make it work.
Many thanks again
Cheers
Rayner
[image: run2.jpg]
[image: run1.jpg]
Dr. Rayner Núñez
Postdoc, Lepidoptera & Trichoptera Section
Centre for Taxonomy and Morphology
***@***.***, ***@***.*** ***@***.***>
…On Mon, 24 Jun 2024 at 11:14, Vincent Bonhomme ***@***.***> wrote:
Hi there,
I don't think that's a matter of shape complexity but how the masks were
done. The current algorithm is only fully reproducible when the masks are
on the middle point of the jpg. If not, the algo will try to find a black
pixel, randomly, on the jpg. Also when single pixels are isolated on the
diagonal eg the top right 'x' below:
```
x o x
x x o
x o o
```
outline extract may/will fail
hope this helps,
if not, send me the masks that produce different outlines and I gonna have
a look,
all the best,
v
Le lun. 24 juin 2024 à 10:59, raynernunez ***@***.***> a
écrit :
> Hi, I'm trying to compare these outlines, Lepidoptera genitalia, but for
> some reason momocs gets wrong, and randomly, some of the images. I
removed
> part of the structure in some samples to compare and the outlines were
> captured fine, look at the two marked samples. Maybe these shapes are
too
> complex? Any help would be great. Thanks.
> body.only.jpg (view on web)
> <
https://github.com/MomX/Momocs/assets/97603334/4b832afb-f285-4f95-8d5a-a32cda647f05>
> complete.jpg (view on web)
> <
https://github.com/MomX/Momocs/assets/97603334/3180c57e-fa30-4361-b741-f319a4fb6b1e>
>
> —
> Reply to this email directly, view it on GitHub
> <#235>, or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/ABECQFQBOQES4IJODZZFOG3ZI7NX3AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM3DSNRVGMZDIMQ>
> .
> You are receiving this because you are subscribed to this thread.Message
> ID: ***@***.***>
>
--
vincentbonhomme.fr <http://www.vincentbonhomme.fr>
cevennette.fr <https://www.cevennette.fr> 🌱
—
Reply to this email directly, view it on GitHub
<#235 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AXIU6BVCNIWGMJCZW73XXJ3ZI7WQ7AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGE4DSNBVG4>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
cant see the images, could you please email them to me?
***@***.***
Le lun. 24 juin 2024 à 13:00, raynernunez ***@***.***> a
écrit :
… Hi, when I created the mask I cropped them at the same size and tried them
to be at the center. Now I tried adding a bit more white background around
the shapes but it is still happening and each time I run it gets erroneous
outlines for different samples (attached 2 examples).
Here are the masks so you can try to figure out how to make it work.
Many thanks again
Cheers
Rayner
[image: run2.jpg]
[image: run1.jpg]
Dr. Rayner Núñez
Postdoc, Lepidoptera & Trichoptera Section
Centre for Taxonomy and Morphology
***@***.***, ***@***.*** ***@***.***>
On Mon, 24 Jun 2024 at 11:14, Vincent Bonhomme ***@***.***>
wrote:
> Hi there,
>
> I don't think that's a matter of shape complexity but how the masks were
> done. The current algorithm is only fully reproducible when the masks
are
> on the middle point of the jpg. If not, the algo will try to find a
black
> pixel, randomly, on the jpg. Also when single pixels are isolated on the
> diagonal eg the top right 'x' below:
>
> ```
> x o x
> x x o
> x o o
> ```
> outline extract may/will fail
>
> hope this helps,
> if not, send me the masks that produce different outlines and I gonna
have
> a look,
>
> all the best,
>
> v
>
>
>
>
> Le lun. 24 juin 2024 à 10:59, raynernunez ***@***.***> a
> écrit :
>
> > Hi, I'm trying to compare these outlines, Lepidoptera genitalia, but
for
> > some reason momocs gets wrong, and randomly, some of the images. I
> removed
> > part of the structure in some samples to compare and the outlines were
> > captured fine, look at the two marked samples. Maybe these shapes are
> too
> > complex? Any help would be great. Thanks.
> > body.only.jpg (view on web)
> > <
>
https://github.com/MomX/Momocs/assets/97603334/4b832afb-f285-4f95-8d5a-a32cda647f05>
>
> > complete.jpg (view on web)
> > <
>
https://github.com/MomX/Momocs/assets/97603334/3180c57e-fa30-4361-b741-f319a4fb6b1e>
>
> >
> > —
> > Reply to this email directly, view it on GitHub
> > <#235>, or unsubscribe
> > <
>
https://github.com/notifications/unsubscribe-auth/ABECQFQBOQES4IJODZZFOG3ZI7NX3AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM3DSNRVGMZDIMQ>
>
> > .
> > You are receiving this because you are subscribed to this
thread.Message
> > ID: ***@***.***>
> >
>
>
> --
> vincentbonhomme.fr <http://www.vincentbonhomme.fr>
> cevennette.fr <https://www.cevennette.fr> 🌱
>
> —
> Reply to this email directly, view it on GitHub
> <#235 (comment)>, or
> unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AXIU6BVCNIWGMJCZW73XXJ3ZI7WQ7AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGE4DSNBVG4>
> .
> You are receiving this because you authored the thread.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#235 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/ABECQFU4IKA75WESSOWUFA3ZI736LAVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGI4TCOJXGA>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
vincentbonhomme.fr <http://www.vincentbonhomme.fr>
cevennette.fr <https://www.cevennette.fr> 🌱
|
|
Hi, thanks, just did
Dr. Rayner Núñez
Postdoc, Lepidoptera & Trichoptera Section
Centre for Taxonomy and Morphology
***@***.***, ***@***.*** ***@***.***>
On Mon, 24 Jun 2024 at 12:12, Vincent Bonhomme ***@***.***>
wrote:
… cant see the images, could you please email them to me?
***@***.***
Le lun. 24 juin 2024 à 13:00, raynernunez ***@***.***> a
écrit :
> Hi, when I created the mask I cropped them at the same size and tried
them
> to be at the center. Now I tried adding a bit more white background
around
> the shapes but it is still happening and each time I run it gets
erroneous
> outlines for different samples (attached 2 examples).
>
> Here are the masks so you can try to figure out how to make it work.
>
> Many thanks again
>
>
> Cheers
>
>
> Rayner
> [image: run2.jpg]
> [image: run1.jpg]
> Dr. Rayner Núñez
>
> Postdoc, Lepidoptera & Trichoptera Section
>
> Centre for Taxonomy and Morphology
> ***@***.***, ***@***.*** ***@***.***>
>
>
>
>
>
> On Mon, 24 Jun 2024 at 11:14, Vincent Bonhomme ***@***.***>
> wrote:
>
> > Hi there,
> >
> > I don't think that's a matter of shape complexity but how the masks
were
> > done. The current algorithm is only fully reproducible when the masks
> are
> > on the middle point of the jpg. If not, the algo will try to find a
> black
> > pixel, randomly, on the jpg. Also when single pixels are isolated on
the
> > diagonal eg the top right 'x' below:
> >
> > ```
> > x o x
> > x x o
> > x o o
> > ```
> > outline extract may/will fail
> >
> > hope this helps,
> > if not, send me the masks that produce different outlines and I gonna
> have
> > a look,
> >
> > all the best,
> >
> > v
> >
> >
> >
> >
> > Le lun. 24 juin 2024 à 10:59, raynernunez ***@***.***> a
> > écrit :
> >
> > > Hi, I'm trying to compare these outlines, Lepidoptera genitalia, but
> for
> > > some reason momocs gets wrong, and randomly, some of the images. I
> > removed
> > > part of the structure in some samples to compare and the outlines
were
> > > captured fine, look at the two marked samples. Maybe these shapes
are
> > too
> > > complex? Any help would be great. Thanks.
> > > body.only.jpg (view on web)
> > > <
> >
>
https://github.com/MomX/Momocs/assets/97603334/4b832afb-f285-4f95-8d5a-a32cda647f05>
>
> >
> > > complete.jpg (view on web)
> > > <
> >
>
https://github.com/MomX/Momocs/assets/97603334/3180c57e-fa30-4361-b741-f319a4fb6b1e>
>
> >
> > >
> > > —
> > > Reply to this email directly, view it on GitHub
> > > <#235>, or unsubscribe
> > > <
> >
>
https://github.com/notifications/unsubscribe-auth/ABECQFQBOQES4IJODZZFOG3ZI7NX3AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM3DSNRVGMZDIMQ>
>
> >
> > > .
> > > You are receiving this because you are subscribed to this
> thread.Message
> > > ID: ***@***.***>
> > >
> >
> >
> > --
> > vincentbonhomme.fr <http://www.vincentbonhomme.fr>
> > cevennette.fr <https://www.cevennette.fr> 🌱
> >
> > —
> > Reply to this email directly, view it on GitHub
> > <#235 (comment)>,
or
> > unsubscribe
> > <
>
https://github.com/notifications/unsubscribe-auth/AXIU6BVCNIWGMJCZW73XXJ3ZI7WQ7AVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGE4DSNBVG4>
>
> > .
> > You are receiving this because you authored the thread.Message ID:
> > ***@***.***>
> >
>
> —
> Reply to this email directly, view it on GitHub
> <#235 (comment)>, or
> unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/ABECQFU4IKA75WESSOWUFA3ZI736LAVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGI4TCOJXGA>
> .
> You are receiving this because you commented.Message ID:
> ***@***.***>
>
--
vincentbonhomme.fr <http://www.vincentbonhomme.fr>
cevennette.fr <https://www.cevennette.fr> 🌱
—
Reply to this email directly, view it on GitHub
<#235 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AXIU6BVFZVV7HTRHQWFATDLZI75KLAVCNFSM6AAAAABJZMNB4KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCOBWGMZDCMJUGM>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi, I'm trying to compare these outlines, Lepidoptera genitalia, but for some reason momocs gets wrong, and randomly, some of the images. I removed part of the structure in some samples to compare and the outlines were captured fine, look at the two marked samples. I ran it several times, each time different outlines were modified. Maybe these shapes are too complex? Any help would be great. Thanks.
The text was updated successfully, but these errors were encountered: