Target detection model optimization method and device based on active learning

The invention discloses a target detection model optimization method and device based on active learning, and the method comprises the steps: firstly, carrying out the detection of unmarked image data through an existing target detection model, selecting the first k samples with the minimum confidence coefficient as candidate samples, extracting image embedding through other pre-training models, and carrying out the detection of a target detection model. Calculating the uniqueness of the model in the candidate samples, selecting the first B samples with the maximum uniqueness so as to enrich the diversity of the candidate set, manually marking the selected first B samples, adding the newly marked samples into the marked sample set, and retraining the model until the performance of the model is not improved any more or the maximum number of iterations is reached; the innovative method not only effectively reduces the marking cost, but also improves the performance of the target detection model, and provides a