Labeling Your Data¶
Labeling sensor data in your application is a critical and often non-recognized part of building a smart application. If you were to follow bad labeling methods it will cascade through your project, causing wasted time and critical classification errors in your end product.
Next, we will go over how the Data Capture Lab makes visualizing and labeling sensor data easy. Note that In the free trial you cannot save changes to projects.
Viewing your capture files¶
After opening your project, click Switch Modes and open Label Explorer mode
To open the raw sensor data files you have captured, use the Project Explorer. This can be found in the upper left-hand corner of the DCL.
Double-Click on a file name to open the file
Depending on the project you have loaded, the file will open and look similar to this
If you are working with a project that has Accelerometer and Gyroscope data, you will notice the DCL has split the data into two tracks. These tracks can be customized to display as many sensor columns as you need.
(Optional) Click on the track menu options to add new tracks, update columns, or remove tracks. This is a powerful feature that allows you to visualize your data easier.
(Optional) Click and drag the area between tracks to adjust a track height
Segments are how we label an event in the DCL. If you are following along in the slide demo, open any of the provided files and you will see a pair of blue and red lines. This defines where our event is located and is called a segment. See screenshot below:
A session separates your labeled events (segments) into a group. Every segment is inside a session. Sessions allow you to work on multiple branches of the same dataset. Sessions can be useful when collaborating with a team or if you want to experiment with different segmentation methods on the same dataset.
Adding Segments to a Session
Select a manual session to work in. In the slide demo, we have (M) My Training Session already created
Open a file
Next, identify your event of interest in the graph. To do this, move your mouse onto the graph of your sensor data and right-click + drag your mouse over the area that you want to label as an event. This will place a new segment in the file.
After creating the segment, you label this segment by clicking Label Current Segment.
Next select the label and click Done
You have now added a new segment to the session My Training Session. The slide demo project has already been labeled, but if you were using your own data set you would repeat this process to label all of the events you wish to use as training data in your project.