Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces CellPrior-Net (CP Net), a novel and efficient pipeline for nuclei detection and classification in hematoxylin and eosin (H&E) whole-slide images, addressing the challenges posed by variations in nuclei morphology and WSI artifacts. By leveraging a lightweight convolutional neural network architecture and utilizing the hematoxylin channel as prior information, CP Net enhances nuclei-aware feature learning while significantly reducing inference time compared to existing methods. Extensive benchmarking across 10.4 million nuclei from diverse datasets demonstrates that CP Net achieves performance on par with state-of-the-art approaches, making it a scalable solution for computational pathology applications.
Achieving state-of-the-art nuclei detection accuracy while slashing inference time, CellPrior-Net redefines efficiency in computational pathology.
Accurate nuclei detection and classification in hematoxylin and eosin (H and E) whole-slide images (WSIs) is a key task in computational pathology, particularly for quantitative analysis of the tumor microenvironment. However, this task remains highly challenging due to variations in nuclei morphology, staining procedures, scanners, organs, magnifications, and WSI artifacts. In addition, many existing pipelines rely on computationally demanding architectures and post-processing procedures, making gigapixel WSI analysis time consuming. In this work, CellPriorNet (CP Net) is proposed, an efficient nuclei detection and classification pipeline that utilizes a lightweight convolutional neural network architecture and hematoxylin (H) channel as prior information to enhance nuclei-aware feature learning. Extensive benchmarking was conducted against state of the art pipelines on 8 public and private datasets (total:10.4M nuclei) obtained from different organs, scanners, magnifications, and clinical centers. Experimental results demonstrate that CP Net achieves comparable performance while significantly reducing inference time. Furthermore, CellQuant Net was introduced, an end to end nuclei quantification pipeline, that integrates a quality assessment (QA) model to exclude regions with artifacts, followed by CP-Net cell detection and classification. The pipeline is publicly available on GitHub, and provides a potentially efficient and scalable framework for downstream computational pathology applications.