Learn More
Œe success of deep learning depends on €nding an architecture to €t the task. As deep learning has scaled up to more challenging tasks, the architectures have become dicult to design by hand. Œis paper proposes an automated method, CoDeepNEAT, for optimizing deep learning architectures through evolution. By extending existing neuroevolution methods to(More)
We present an information hiding technique that utilizes lifting schemes to effectively hide information in images. A successful information hiding should result in the extraction of the hidden data from the image with high degree of data integrity. Current trends favor using digital image files as the cover file to hide another digital file that contains(More)
While the robot is inmotion, path planning should follow the three aspects: (1) acquire the knowledge from its environmental conditions. (2) determine its position in the environment and (3) decision-making and execution to achieve its highest-order goals. The present research work aims to develop an efficient particle swarm optimizationbased path planner(More)
  • 1