Using any current state of the art object detection model requires a large amount of labeled data that might be very expensive and time consuming to acquire. It is hard to keep track of the model’s performance after it is in production. In addition to that, the model should be retrained and updated with the data that it has seen during production.
In order to cut labeling costs and achieve high performance while selecting only valuable data we use active learning where the learning algorithms can actively teach itself. This type of iterative supervised learning lets the model choose from the data based on its performance on the samples.
At IxorThink a simple framework for using active learning in object detection has been proposed to significantly save time and resources on the labeling efforts to train an object detection model. In addition to that, the proposed work proved to be effective in improving the model’s performance after the production stage.
For a more in-dept explanation, please take a look at our Medium blog.