Application of neural nets to seismic signal analysis

We present a comparative study of the performance of reported neural net algorithms for the detection of first breaks in seismic reflection data with regard to accuracy, learning rate and generalisability. In addition we suggest a new approach that produces improved results.<
>