DyRA: Portable Dynamic Resolution Adjustment Network for Existing Detectors
This addresses the problem of object size variability for computer vision practitioners by offering a portable solution to enhance existing detectors, though it is incremental as it builds on multi-resolution strategies.
The paper tackles the challenge of maintaining constant accuracy in object detection across varying object sizes by introducing DyRA, a dynamic resolution adjustment network that provides image-specific scale factors for existing detectors, improving accuracy across models like RetinaNet and Faster-RCNN.
Achieving constant accuracy in object detection is challenging due to the inherent variability of object sizes. One effective approach to this problem involves optimizing input resolution, referred to as a multi-resolution strategy. Previous approaches to resolution optimization have often been based on pre-defined resolutions with manual selection. However, there is a lack of study on run-time resolution optimization for existing architectures. This paper introduces DyRA, a dynamic resolution adjustment network providing an image-specific scale factor for existing detectors. This network is co-trained with detectors utilizing specially designed loss functions, namely ParetoScaleLoss and BalanceLoss. ParetoScaleLoss determines an adaptive scale factor for robustness, while BalanceLoss optimizes overall scale factors according to the localization performance of the detector. The loss function is devised to minimize the accuracy drop across contrasting objectives of different-sized objects for scaling. Our proposed network can improve accuracy across various models, including RetinaNet, Faster-RCNN, FCOS, DINO, and H-Deformable-DETR. The code is available at https://github.com/DaEunFullGrace/DyRA.git.