IAUNet: Instance-Aware U-Net

2025-08-03

AI Summary

To address the challenging instance segmentation of overlapping and multi-scale cells in biomedical images, this paper proposes IAUNet—the first lightweight, end-to-end framework that deeply integrates a U-Net encoder with a query-based mechanism. Methodologically, IAUNet introduces a lightweight convolutional Pixel Decoder and a multi-scale Transformer Query Decoder to enable efficient instance-aware decoding within the U-Net feature space; it further incorporates cross-scale feature fusion and object queries to significantly improve separation accuracy for highly overlapping cells. As a key contribution, we construct and publicly release Revvity 2025, a high-quality, fully annotated cell segmentation dataset. Extensive experiments demonstrate that IAUNet consistently outperforms state-of-the-art CNN-, Transformer-, and query-based methods across multiple public benchmarks and Revvity 2025, establishing a new performance baseline for cell instance segmentation.

Technology Category

• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Instance segmentation is critical in biomedical imaging to accurately distinguish individual objects like cells, which often overlap and vary in size. Recent query-based methods, where object queries guide segmentation, have shown strong performance. While U-Net has been a go-to architecture in medical image segmentation, its potential in query-based approaches remains largely unexplored. In this work, we present IAUNet, a novel query-based U-Net architecture. The core design features a full U-Net architecture, enhanced by a novel lightweight convolutional Pixel decoder, making the model more efficient and reducing the number of parameters. Additionally, we propose a Transformer decoder that refines object-specific features across multiple scales. Finally, we introduce the 2025 Revvity Full Cell Segmentation Dataset, a unique resource with detailed annotations of overlapping cell cytoplasm in brightfield images, setting a new benchmark for biomedical instance segmentation. Experiments on multiple public datasets and our own show that IAUNet outperforms most state-of-the-art fully convolutional, transformer-based, and query-based models and cell segmentation-specific models, setting a strong baseline for cell instance segmentation tasks. Code is available at https://github.com/SlavkoPrytula/IAUNet

Problem

• Improves instance segmentation of overlapping biomedical objects like cells
• Enhances U-Net efficiency with lightweight Pixel decoder and Transformer decoder
• Introduces new dataset for benchmarking cell instance segmentation

Innovation

• Lightweight convolutional Pixel decoder enhances efficiency
• Transformer decoder refines multi-scale object features
• Query-based U-Net architecture for instance segmentation