SUPA BOLT Image Classification 101

Learn about an overview of image classification, its use cases and how you can get started with image classification on SUPA BOLT. 

Image classification is a data labeling task for assigning a label or class to an entire image. Image classification models take an image as input and return a prediction about which classes the image belongs to.

Use Cases

Image classification is used when we are not interested in specific instances of objects in an image with location or spatial information. The use cases involve matching visual content to predefined categories. Example use cases:

Keyword Classification

Assign keywords to each image in stock photography

Image Search

Organize and categorize photo galleries using keywords or tags for better browsing experience

Image Moderation

Detect and filter any images with undesired content such as gore

Recommendation System

Recommend similar products based on features of an image

Image Annotation Validation

Select Approve or Reject for an image with completed annotations to validate quality

‍Image Classification with SUPA BOLT

In this section below, we'll cover the steps to launching your image classification project.

Step 1: Create a new image classification project 

Step 2: Upload image data, either via local or S3

Step 3: Add questions and define the categories for the labelers to choose from

For detailed steps on setting up taxonomy/questions, click here.

Step 4: Start your project. It’s good to go!

‍Conclusion

Image classification can be considered the fundamental problem in the field of computer vision. It doesn’t require detecting the objects in the image, simply classifying the image itself. 

Launch your image classification project on SUPA BOLT today!