Computer Vision News - May 2020

2 Summary Artificial Intelligence 30 Computer Vision News has found great new stories, written somewhere else by somebody else. We share them with you, adding a short comment. Enjoy! A I S P O T L I G H T N E W S m y Artifi i l I t l i 4 CORaiL: Using Artificial Intelligence to Save Coral Reefs Let’s start by honoring Earth Day 2020! Coral reefs are among the world's most precious ecosystems, providing habitat and shelter for approximately 25% of global marine life. As of today, they are heavily endangered by warming temperatures, unsustainable coastal development, overfishing and a terrible thing called bottom trawling. CORaiL is a joint project by Accenture, Intel and the Sulubaaï Environmental Foundation, using Artificial Intelligence in a fascinating new way to help saving the Coral Reef . It was first deployed to the reef surrounding Pangatalan Island in the Philippines. Read More SAS calls for crowd-driven artificial intelligence to help track deforestation AI is not only called to safe seabed! SAS (the software firm, not the airline) wants to implement the next generation of crowd-driven artificial intelligence and scientific analysis to help us better understand our planet. All kinds of volunteers are called to kick-start this effort by reviewing and judging images of the rainforest . Their contribution will be used to develop AI models that will exponentially increase the value of human insights and strive to deliver near real- time assessment of global environmental change , thus helping to drive vital policy responses to protect forests more quickly. Read More CV and DL Techniques for the Analysis of Drone-Acquired Forest Images This very recently published (and lovely) paper is a transfer learning study, by which researchers study and quantify issues related to the use of Deep Learning with their own UAV-acquired images in forestry applications. The purpose of the model is to acquire detailed knowledge about mixed forest structure and composition, in order to properly understand current status and future interventions. This work proves the development possibilities of Deep Learning to solve practical problems in forestry research. It also shows transfer learning as essential to obtain good results for patch classification. Read More