Learning Dynamic Query Combinations for Transformer-based Object Detection and Segmentation

TL;DR

This paper introduces a method to generate dynamic queries for transformer-based detection and segmentation by learning convex combinations of queries, improving performance across various tasks.

Contribution

It proposes a novel approach to create modulated queries with dynamic coefficients, enhancing the ability of DETR-based models to capture object priors and improve accuracy.

Findings

Consistent performance improvements across multiple tasks.

Dynamic queries outperform static learned queries.

Applicable to various transformer-based detection models.

Abstract

Transformer-based detection and segmentation methods use a list of learned detection queries to retrieve information from the transformer network and learn to predict the location and category of one specific object from each query. We empirically find that random convex combinations of the learned queries are still good for the corresponding models. We then propose to learn a convex combination with dynamic coefficients based on the high-level semantics of the image. The generated dynamic queries, named modulated queries, better capture the prior of object locations and categories in the different images. Equipped with our modulated queries, a wide range of DETR-based models achieve consistent and superior performance across multiple tasks including object detection, instance segmentation, panoptic segmentation, and video instance segmentation.