Detecting the open-world objects with the help of the Brain

TL;DR

This paper introduces a novel open-world object detection method that uses large pre-trained vision-language models as a 'brain' to identify unknown objects, employing a special loss function and pseudo-labeling to improve detection and learning.

Contribution

It proposes leveraging large pre-trained vision-language models as a 'brain' for OWOD, with new loss functions and a pseudo-labeling scheme for better unknown object detection.

Findings

Effective detection of unknown objects in open-world scenarios.

Improved incremental learning of novel objects.

Utilization of VL models enhances open-world detection capabilities.

Abstract

Open World Object Detection (OWOD) is a novel computer vision task with a considerable challenge, bridging the gap between classic object detection (OD) benchmarks and real-world object detection. In addition to detecting and classifying seen/known objects, OWOD algorithms are expected to detect unseen/unknown objects and incrementally learn them. The natural instinct of humans to identify unknown objects in their environments mainly depends on their brains' knowledge base. It is difficult for a model to do this only by learning from the annotation of several tiny datasets. The large pre-trained grounded language-image models - VL (\ie GLIP) have rich knowledge about the open world but are limited to the text prompt. We propose leveraging the VL as the ``Brain'' of the open-world detector by simply generating unknown labels. Leveraging it is non-trivial because the unknown labels impair the model's learning of known objects. In this paper, we alleviate these problems by proposing the down-weight loss function and decoupled detection structure. Moreover, our detector leverages the ``Brain'' to learn novel objects beyond VL through our pseudo-labeling scheme.