Session expiration Your session is going to expireClick here to extend


Small project <800

Posted on

4/15/19 6:13 PM



This project is currently receiving offers

Post a similar project and get non-binding offers in no time.

Post similar project now


I have a test and measurement device that measures a pipeline's internal diameter. The pattern of change in the Internal diameter corresponds to a pipeline feature. I have thousands of datasets where the manual analysis is performed and the features identified and reported. Although the actual scope of my work involves detection / identification and measurement of these features, I am only interested in detection / identification at this point. I am attaching our standard deliverable which is the list of features identified in the pipeline. The dataset itself is a collection of binary files grouped into 1000 feet of pipeline data. Depending on the length of the pipeline data, some times there is just one file in the dataset and some times there are thousands. I can provide library functions to read the dataset itself. What I need is a training engine that I can feed my datasets to (in any format you want) and the output spreadsheet that was created manually using the dataset. The goal is to feed the new datasets to the training model and the engine outputting the list of features.