Domain adaptation using pseudo-labelling and model certainty quantification for video data

Abstract

Systems and method for domain adaptation using pseudo-labelling and model certainty quantification for video data are provided. The method includes obtaining a source data and a target data each comprising a plurality of frames for processing by a machine learning module. The method comprises testing the target data to identify if a minimum number of frames exhibit a frame confidence score based on the source data and identifying salient region within the target data and measuring a degree of spatial consistency of the salient region over time. The method comprises identifying class specific attention region within the target data and measuring a confidence score of class specific attention region within the target data and carrying out pseudo-labeling of the target data based on the source data and calculating a certainty metrics value based on the frame confidence score, the degree of spatial consistency of the salient region over time, the confidence score of class specific attention region within the frames of the target data and confidence score of the pseudo-labeling on the target data. The machine learning module is retrained till the certainty metrics value reaches peak and further retraining the machine learning module does not increase the certainty metrics value.