OmniLabel: A Challenging Benchmark for Language-Based Object Detection
This provides a challenging benchmark for researchers in computer vision and natural language processing to advance language-based object detection, though it is incremental as it builds on existing evaluation needs.
The authors tackled the lack of proper evaluation for language-based object detection by introducing OmniLabel, a benchmark with a novel task definition, dataset of over 28K unique object descriptions on 25K images, and a modified average precision metric, which they validated by evaluating strong baselines.
Language-based object detection is a promising direction towards building a natural interface to describe objects in images that goes far beyond plain category names. While recent methods show great progress in that direction, proper evaluation is lacking. With OmniLabel, we propose a novel task definition, dataset, and evaluation metric. The task subsumes standard- and open-vocabulary detection as well as referring expressions. With more than 28K unique object descriptions on over 25K images, OmniLabel provides a challenging benchmark with diverse and complex object descriptions in a naturally open-vocabulary setting. Moreover, a key differentiation to existing benchmarks is that our object descriptions can refer to one, multiple or even no object, hence, providing negative examples in free-form text. The proposed evaluation handles the large label space and judges performance via a modified average precision metric, which we validate by evaluating strong language-based baselines. OmniLabel indeed provides a challenging test bed for future research on language-based detection.