Parameter-Inverted Image Pyramid Networks

TL;DR

The paper introduces Parameter-Inverted Image Pyramid Networks (PIIP), a novel architecture that processes multi-scale images with varying model sizes to improve efficiency and performance in vision tasks.

Contribution

PIIP employs models with different parameter sizes for each resolution level and introduces a feature interaction mechanism, reducing computation while enhancing multi-scale feature integration.

Findings

Achieves superior performance in object detection, segmentation, and classification.

Reduces computational cost by 40-60% compared to traditional methods.

Improves large-scale vision model performance by 1-2%.

Abstract

Image pyramids are commonly used in modern computer vision tasks to obtain multi-scale features for precise understanding of images. However, image pyramids process multiple resolutions of images using the same large-scale model, which requires significant computational cost. To overcome this issue, we propose a novel network architecture known as the Parameter-Inverted Image Pyramid Networks (PIIP). Our core idea is to use models with different parameter sizes to process different resolution levels of the image pyramid, thereby balancing computational efficiency and performance. Specifically, the input to PIIP is a set of multi-scale images, where higher resolution images are processed by smaller networks. We further propose a feature interaction mechanism to allow features of different resolutions to complement each other and effectively integrate information from different spatial scales. Extensive experiments demonstrate that the PIIP achieves superior performance in tasks such as object detection, segmentation, and image classification, compared to traditional image pyramid methods and single-branch networks, while reducing computational cost. Notably, when applying our method on a large-scale vision foundation model InternViT-6B, we improve its performance by 1%-2% on detection and segmentation with only 40%-60% of the original computation. These results validate the effectiveness of the PIIP approach and provide a new technical direction for future vision computing tasks. Our code and models are available at https://github.com/OpenGVLab/PIIP.