Trjtdtk

slides for 30min~1hr skype tenure track application interview

"as much details as you can remember"

How to check whether the reindex working or not in Magento?

Does a dangling wire really electrocute me if I'm standing in water?

The phrase "to the numbers born"?

APIPA and LAN Broadcast Domain

For what reasons would an animal species NOT cross a *horizontal* land bridge?

Are spiders unable to hurt humans, especially very small spiders?

Can withdrawing asylum be illegal?

Right tool to dig six foot holes?

Vorinclex, does my opponents land untap if they were tapped before i summoned him?

How to translate "being like"?

If I score a critical hit on an 18 or higher, what are my chances of getting a critical hit if I roll 3d20?

Why doesn't shell automatically fix "useless use of cat"?

I am eight letters word. Find me who Am I?

Did any laptop computers have a built-in 5 1/4 inch floppy drive?

When should I buy a clipper card after flying to Oakland?

Balance problems for leveling up mid-fight?

What does Linus Torvalds mean when he says that Git "never ever" tracks a file?

What could be the right powersource for 15 seconds lifespan disposable giant chainsaw?

Can an undergraduate be advised by a professor who is very far away?

Why is the Constellation's nose gear so long?

How do I free up internal storage if I don't have any apps downloaded?

Why doesn't mkfifo with a mode of 1755 grant read permissions and sticky bit to the user?

How to map RGB image segmentation ground truth to classes/one-hot vectors in TensorFlow?

0

$begingroup$

Is there a TensorFlow function which takes multicategorical ground truth for semantic image segmentations in the form of RGB images and outputs a tensor with a one-hot vector encoding of the corresponding class per pixel? In other words, I'm starting with RGB images for the ground truth, so each class has a distinct color, like this:

This image was taken from the SYNTHIA dataset.Many semantic segmentation datasets supply their ground truth this way. Every pixel has just one class. What I am looking for is a function that first enumerates the number of different colours in an annotation, and then considers each colour to be a different class automatically. I thought that usually when performing semantic segmentation, each ground truth class is encoded using a one hot vector, to which predicted class probabilities can easily be compared.

I'm basically asking the exact same as this question, and am merely wondering whether such a function has been added since that question was posed, or if someone has a more efficient solution. The answer to it seems convoluted, and I can't imagine such functionality does not exist, as it would seem like a common task. Also, while for many datasets the ground truth is additionally given as some sort of text file (like JSON), writing parser for each different dataset you use seems needlessly cumbersome.

python tensorflow dataset annotation

share|improve this question

asked Mar 31 at 10:51

EmielBossEmielBoss

185

$endgroup$

add a comment | 

0

$begingroup$

Is there a TensorFlow function which takes multicategorical ground truth for semantic image segmentations in the form of RGB images and outputs a tensor with a one-hot vector encoding of the corresponding class per pixel? In other words, I'm starting with RGB images for the ground truth, so each class has a distinct color, like this:

This image was taken from the SYNTHIA dataset.Many semantic segmentation datasets supply their ground truth this way. Every pixel has just one class. What I am looking for is a function that first enumerates the number of different colours in an annotation, and then considers each colour to be a different class automatically. I thought that usually when performing semantic segmentation, each ground truth class is encoded using a one hot vector, to which predicted class probabilities can easily be compared.

I'm basically asking the exact same as this question, and am merely wondering whether such a function has been added since that question was posed, or if someone has a more efficient solution. The answer to it seems convoluted, and I can't imagine such functionality does not exist, as it would seem like a common task. Also, while for many datasets the ground truth is additionally given as some sort of text file (like JSON), writing parser for each different dataset you use seems needlessly cumbersome.

python tensorflow dataset annotation

share|improve this question

asked Mar 31 at 10:51

EmielBossEmielBoss

185

$endgroup$

add a comment | 

$begingroup$

Is there a TensorFlow function which takes multicategorical ground truth for semantic image segmentations in the form of RGB images and outputs a tensor with a one-hot vector encoding of the corresponding class per pixel? In other words, I'm starting with RGB images for the ground truth, so each class has a distinct color, like this:

This image was taken from the SYNTHIA dataset.Many semantic segmentation datasets supply their ground truth this way. Every pixel has just one class. What I am looking for is a function that first enumerates the number of different colours in an annotation, and then considers each colour to be a different class automatically. I thought that usually when performing semantic segmentation, each ground truth class is encoded using a one hot vector, to which predicted class probabilities can easily be compared.

I'm basically asking the exact same as this question, and am merely wondering whether such a function has been added since that question was posed, or if someone has a more efficient solution. The answer to it seems convoluted, and I can't imagine such functionality does not exist, as it would seem like a common task. Also, while for many datasets the ground truth is additionally given as some sort of text file (like JSON), writing parser for each different dataset you use seems needlessly cumbersome.

python tensorflow dataset annotation

share|improve this question

asked Mar 31 at 10:51

EmielBossEmielBoss

185

$endgroup$

share|improve this question

asked Mar 31 at 10:51

EmielBossEmielBoss

185

share|improve this question

asked Mar 31 at 10:51

EmielBossEmielBoss

185

asked Mar 31 at 10:51

EmielBossEmielBoss

185

asked Mar 31 at 10:51

EmielBossEmielBoss

185