Abstract

Classification of acoustic or seismic images is a highly challenging and computationally intensive and efficacy of any algorithm depends on the order of pre-processing and quality of images. Artificial intelligence techniques based on Machine Language and Neural Networks found to be very promising in various fields of engineering and science applications. Deep learning using Convolutional Neural Networks (CNN), a subset of AI reported to be promising and enhances the efficacy of classification. The present study deals with application of CNN to demonstrate its potential in classification of subbottom seismic/acoustic images associated even with a low contrast in their texture. The results are quite promising and CNN classifies over 75 % of images correctly even with limited data used in training of the network. The results can be improved with more number of images data sets in training.