InternImage: Exploring Large-Scale Vision Foundation Models with Deformable Convolutions

TL;DR

InternImage introduces a large-scale CNN model utilizing deformable convolutions, achieving state-of-the-art results on benchmarks like COCO and ADE20K, and bridging the performance gap with vision transformers.

Contribution

The paper presents a novel large-scale CNN model, InternImage, that leverages deformable convolutions to enhance receptive fields and adaptively learn robust patterns from massive data.

Findings

Achieved 65.4 mAP on COCO test-dev

Reached 62.9 mIoU on ADE20K

Outperformed existing CNNs and ViTs on key benchmarks

Abstract

Compared to the great progress of large-scale vision transformers (ViTs) in recent years, large-scale models based on convolutional neural networks (CNNs) are still in an early state. This work presents a new large-scale CNN-based foundation model, termed InternImage, which can obtain the gain from increasing parameters and training data like ViTs. Different from the recent CNNs that focus on large dense kernels, InternImage takes deformable convolution as the core operator, so that our model not only has the large effective receptive field required for downstream tasks such as detection and segmentation, but also has the adaptive spatial aggregation conditioned by input and task information. As a result, the proposed InternImage reduces the strict inductive bias of traditional CNNs and makes it possible to learn stronger and more robust patterns with large-scale parameters from massive data like ViTs. The effectiveness of our model is proven on challenging benchmarks including ImageNet, COCO, and ADE20K. It is worth mentioning that InternImage-H achieved a new record 65.4 mAP on COCO test-dev and 62.9 mIoU on ADE20K, outperforming current leading CNNs and ViTs. The code will be released at https://github.com/OpenGVLab/InternImage.