dc.description.abstract | In order to overcome the problem that the traditional resistivity images is challenging to be automation due to the long time consuming and anthropogenic influence, there is a proposal that package of resistivity images based on the Deep Neural Network can result in the positive neural network, which significantly shortens the time of the resistivity images. The positive neural network could be images quickly. However, the whole training is lengthy, and the result might be disappointing; even though renewing the process by generating the new training data and changing parameters is not only complicated but also not knowing the finish time, and the predicted outcomes have to be revealed until on the device, which has the function of the Graphics Processing Unit (GPU). For the purpose of solving the problem as mentioned above, we send the parameters to the Client by rewriting the calculated program with GPU into the Application Programming Interface (API) and setting it up on the device with GPU and then designs an application by Graphical User Interface (GUI) to the Client. As long as there is a network environment, we can receive the data of the training model and previous results, and even to stop or modify the training data and resend the parameters to the application for forecasting from time to time. With the GUI, the technology could increase its penetration of the market sharply. | en_US |