System and method of incremental learning for object detection

Abstract

Methods and systems perform incremental learning object detection in images and / or videos without catastrophic forgetting of previously-learned object classes. A two-stage neural network object detector is trained to locate and identify objects pertaining to an additional object class by iteratively updating the two-stage neural network object detector until an overall detection accuracy criterion is met. The updating is performed so as to balance minimizing a loss of an initial ability to locate and identify objects pertaining to the previously-learned object classes and maximizing an ability to additionally locate and identify the objects pertaining to the additional object class. Assessing whether the overall detection accuracy criterion is met compares outputs of an initial version of the two- stage neural network object detector with a current region proposal output by a current version of the two-stage neural network object detector to determining a region proposal distillation loss and a previously-learned-object identification distillation loss.

Description

technical field [0001] The present invention relates to detecting objects in images and / or videos. Background technique [0002] Various real-world applications such as automated analysis of medical images and image recognition-based secure detection of objects (i.e. instances of known object classes) in images and / or videos. Recently, object detection in images and / or videos has benefited significantly from the development of neural networks. As used herein, the term "one or more object detectors" is a shorter form of "one or more neural network object detectors". Neural network object detectors are essentially models characterized by sets of parameters. Object detection involves determining the location and type or class of objects in images and / or videos. Conventional object detectors utilize one-stage or two-stage object detectors. Single-stage object detectors (e.g., You-Only-Look-Once, YOLO, and Single ShotDetection, SSD) simultaneously determine where and what obj...

AI Technical Summary

Problems solved by technology

The article discusses using knowledge distillation to add new categories to a region classifier, but this is problematic because the above method cannot be directly applied to more advanced architectures such as Faster RCNN

Images