Image segmentation losses with modules expressing a relationship between predictions

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61988987%3A17610%2F22%3AA2302FC5" target="_blank" >RIV/61988987:17610/22:A2302FC5 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1007/978-3-031-08974-9_27" target="_blank" >http://dx.doi.org/10.1007/978-3-031-08974-9_27</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/978-3-031-08974-9_27" target="_blank" >10.1007/978-3-031-08974-9_27</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Image segmentation losses with modules expressing a relationship between predictions

Popis výsledku v původním jazyce

We focus on semantic image segmentation with the usage of deep neural networks and give emphasis on the loss functions used for training the networks. Considering region-based losses, Dice loss, and Tversky loss, we propose two independent modules that easily modify the loss functions to take into account the relationship between the class predictions and increase the slope of the gradient. The first module expresses the ambiguity between classes and the second module utilizes a differentiable soft argmax function. Each of the modules is used before the standard loss is computed and remains untouched. In the benchmark, we involved two neural network architectures with two different backbones, selected two loss functions, and examined separately two scenarios for softmax and sigmoid top activation functions. In the experiment, we demonstrate the usefulness of our modules by improving the IOU and F1 coefficients on the test dataset for all scenarios tested. Moreover, the usage of the modules decreases overfitting. The proposed modules are easy to integrate into existing solutions and add near-zero computational overhead.

Název v anglickém jazyce

Image segmentation losses with modules expressing a relationship between predictions

Popis výsledku anglicky

We focus on semantic image segmentation with the usage of deep neural networks and give emphasis on the loss functions used for training the networks. Considering region-based losses, Dice loss, and Tversky loss, we propose two independent modules that easily modify the loss functions to take into account the relationship between the class predictions and increase the slope of the gradient. The first module expresses the ambiguity between classes and the second module utilizes a differentiable soft argmax function. Each of the modules is used before the standard loss is computed and remains untouched. In the benchmark, we involved two neural network architectures with two different backbones, selected two loss functions, and examined separately two scenarios for softmax and sigmoid top activation functions. In the experiment, we demonstrate the usefulness of our modules by improving the IOU and F1 coefficients on the test dataset for all scenarios tested. Moreover, the usage of the modules decreases overfitting. The proposed modules are easy to integrate into existing solutions and add near-zero computational overhead.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

10102 - Applied mathematics

Projekt

<a href="/cs/project/EF17_049%2F0008414" target="_blank" >EF17_049/0008414: Centrum pro výzkum a vývoj metod umělé intelligence v automobilovém průmyslu regionu</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Information Processing and Management of Uncertainty in Knowledge-Based Systems

ISBN

978-3-031-08973-2

ISSN

1865-0929

e-ISSN

1865-0937

Počet stran výsledku

12

Strana od-do

343-354

Název nakladatele

Springer

Místo vydání

Milano

Místo konání akce

Milano

Datum konání akce

11. 7. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—