scholarly journals The Camouflage Machine: Optimizing protective coloration using deep learning with genetic algorithms

Evolution ◽  
2021 ◽  
Author(s):  
John G. Fennell ◽  
Laszlo Talas ◽  
Roland J. Baddeley ◽  
Innes C. Cuthill ◽  
Nicholas E. Scott‐Samuel

Author(s):  
Hossam Magdy Balaha ◽  
Hesham Arafat Ali ◽  
Esraa Khaled Youssef ◽  
Asmaa Elsayed Elsayed ◽  
Reem Adel Samak ◽  
...  




2019 ◽  
Vol 2019 ◽  
pp. 1-12 ◽  
Author(s):  
Fernando Mattioli ◽  
Daniel Caetano ◽  
Alexandre Cardoso ◽  
Eduardo Naves ◽  
Edgard Lamounier

The choice of a good topology for a deep neural network is a complex task, essential for any deep learning project. This task normally demands knowledge from previous experience, as the higher amount of required computational resources makes trial and error approaches prohibitive. Evolutionary computation algorithms have shown success in many domains, by guiding the exploration of complex solution spaces in the direction of the best solutions, with minimal human intervention. In this sense, this work presents the use of genetic algorithms in deep neural networks topology selection. The evaluated algorithms were able to find competitive topologies while spending less computational resources when compared to state-of-the-art methods.





2020 ◽  
Author(s):  
J. G. Fennell ◽  
L. Talas ◽  
R. J. Baddeley ◽  
I. C. Cuthill ◽  
N. E. Scott-Samuel

AbstractThe essential problem in visual detection is separating an object from its background. Whether in nature or human conflict, camouflage aims to make the problem harder, while conspicuous signals (e.g. for warning or mate attraction) require the opposite. Our goal is to provide a reliable method for identifying the hardest and easiest to find patterns, for any given environment. The problem is challenging because the parameter space provided by varying natural scenes and potential patterns is vast. Here we successfully solve the problem using deep learning with genetic algorithms and illustrate our solution by identifying appropriate patterns in two environments. To show the generality of our approach, we do so for both trichromatic and dichromatic visual systems. Patterns were validated using human participants; those identified as the best camouflage were significantly harder to find than a widely adopted military camouflage pattern, while those identified as most conspicuous were significantly easier than other patterns. Our method, dubbed the ‘Camouflage Machine’, will be a useful tool for those interested in identifying the most effective patterns in a given context.



IEEE Access ◽  
2020 ◽  
Vol 8 ◽  
pp. 100631-100644 ◽  
Author(s):  
Abdullah Asim Yilmaz ◽  
Mehmet Serdar Guzel ◽  
Erkan Bostanci ◽  
Iman Askerzade


Author(s):  
F. Habeeb ◽  
Sherihan Abuelenin ◽  
Samir Elmougy


Sign in / Sign up

Export Citation Format

Share Document