Label Data and Create Datasets

To manage your data you can use labels and datasets.

Label data

To label data, go to the DATA tab in the Viam app.

On the Images subtab, you can filter available images, using the Filtering menu and select the attributes that match where, how, and when the data was collected.

You can label your data with:

• Bounding boxes, which you can use to train object detection models.
• Image tags, which you can use to train image classification models.

Bounding boxes

You can create one or more bounding boxes for objects in each image. If you annotate an entire dataset, you can use these bounding boxes to create object detection models. For example, if you would like to create a model that detects a dog in an image, add bounding boxes around the dog in each of your images and add or select the label dog.

To add a bounding box, click on the image you want to add the bounding box to. More information about the image will open up to the right side. Select the Bounding Box option. Choose an existing label or create a new label. Click on the image where you would like to add the bounding box and drag to where the bounding box should end.

If you want to expand the image, click on the expand side menu arrow in the corner of the image:

Repeat this with all images in your dataset. To see all the images that have bounding boxes, you can filter your dataset by selecting the label from the Bounding box labels dropdown in the Filtering menu.

To delete a bounding box, click the cross next to the corresponding label.

Image tags

You can use tags to create classification models for images. For example, if you would like to create a model that identifies an image of a star in a set of images, tag each image containing a star with a star tag. The filter also needs to include images without the star tag or with another tag like notstar. If you add a notstar tag, you can filter the data in your dataset by selecting star and notstar from the Tags dropdown in the Filtering menu. Alternatively you can use date ranges to filter for relevant data.

To tag an image, click on the image. More information about the image will open up to the right side. Select the Image Tags option.

If you want to expand the image, click on the expand side menu arrow in the corner of the image.

Repeat this with all images in your dataset.

To delete a tag, click the cross next to the corresponding label.

Datasets

Create a dataset and add data

To create a dataset, go to the DATA tab in the Viam app.

On the Images subtab, you can filter available images, using the Filtering menu and select the attributes that match where, how, and when the data was collected.

To add an image to a new dataset, click on the image and create a dataset in the Datasets dropdown. Once you have created a dataset, you can select the existing dataset from the dropdown.

You can also create a new dataset from the DATASETS subtab. Click Create dataset, enter a name or use the suggested name for your dataset and click Create dataset again.

Alternatively, you can use the viam data dataset add command to add an image or group of images matching a specific filter to a dataset using the Viam CLI.

Remove an image from a dataset

To remove an image from a dataset click on the delete button next to the dataset name.

Alternatively, you can use the viam data dataset remove command to remove an image or group of images matching a specific filter using the Viam CLI.

View the data in a dataset

Once you have added data to your dataset, you can view the data in your dataset by clicking on your dataset’s name on the image sidebar or on the DATASETS subtab.

You can click on any image in the dataset and use its appearing menu in order to add or remove bounding box labels or tags, or remove that image from the dataset.

Train a model on a dataset

To train a model on a dataset, see Train a model.

Rename a dataset

To rename a dataset, go to the DATASETS subtab. Then right click on the dataset and click on Rename in the menu that appears, enter a new name and hit enter. Alternatively, you can also click on the dataset and rename it from the three-dot icon next to the dataset name.

Delete a dataset

To delete a dataset, go to the DATASETS subtab, right click on the dataset and click on Delete in the menu that appears. Alternatively, you can also click on the dataset and delete it from the three-dot icon next to the dataset name.

Next steps

Train Model

Train an image classification model on labeled image data.

Smart Pet Treat Dispenser

Use a Raspberry Pi, a motor, and machine learning to build a smart pet feeder.