Fast And Complex Moving Object Detection Using Gan In Computer Vision

  • Mr.R.RAJ BHARATH, S.SOWNDHARYA, SWETHA PRADEEP KOOVA, V.VINOTHINI
Keywords: Generative adversarial network, Video frame interpolation, Frame inclusion.

Abstract

Fast moving object detection and video frame interpolation usually applies convolution neural networks. But most of these models deal with difficulties due to sudden illumination change, occlusion, blur, etc. With the advancement in generative adversarial network a dense attention generative adversarial network that deals with these difficulties by combing the global and local data and providing network accuracy in focusing on moving objects is proposed. Training the neural network using wide range of video data and formulation of video interpolation as a single convolution process allowed eliminate the problems and enables high-quality video frame interpolation. Adversarial learning and GAN work in development of more satisfactory outcomes. We have to converted the video into image framesand detection of fast- moving object is done. Then the selected frames undergo series of operations for feature extraction. Further the frames to be included for video enhancement is processed and included. Finally, the image frames are converted back to video. Experimental results on various datasets ensure that this method has more effectiveoutcome.
Published
2021-09-17
How to Cite
SWETHA PRADEEP KOOVA, V.VINOTHINI, M. B. S. (2021). Fast And Complex Moving Object Detection Using Gan In Computer Vision. Design Engineering, 12445-12454. Retrieved from http://www.thedesignengineering.com/index.php/DE/article/view/4403
Section
Articles