Abstract
Stain normalization is an important processing task for computer-aided diagnosis (CAD) systems in modern digital pathology. This task reduces the color and intensity variations present in stained images from different laboratories. Consequently, stain normalization typically increases the prediction accuracy of CAD systems. However, there are computational challenges that this normalization step must overcome, especially for real-time applications: the memory and run-time bottlenecks associated with the processing of images in high resolution, e.g., 40X. Moreover, stain normalization can be sensitive to the quality of the input images, e.g., when they contain stain spots or dirt. In this case, the algorithm may fail to accurately estimate the stain vectors. We present a high-performance system for stain normalization using a state-of-the-art unsupervised method based on stain-vector estimation. Using a highly-optimized normalization engine, our architecture enables high-speed and large-scale processing of high-resolution whole-slide images. This optimized engine integrates an automated thresholding technique to determine the useful pixels and uses a novel pixel-sampling method that significantly reduces the processing time of the normalization algorithm. We demonstrate the performance of our architecture using measurements from images of different sizes and scanner formats that belong to four different datasets. The results show that our optimizations achieve up to 58x speedup compared to a baseline implementation. We also prove the scalability of our system by showing that the processing time scales almost linearly with the amount of tissue pixels present in the image. Furthermore, we show that the output of the normalization algorithm can be adversely affected when the input images include artifacts. To address this issue, we enhance the stain normalization pipeline by introducing a parameter cross-checking technique that automatically detects the distortion of the algorithm's critical parameters. To assess the robustness of the proposed method we employ a machine learning (ML) pipeline that classifies images for detection of prostate cancer. The results show that the enhanced normalization algorithm increases the classification accuracy of the ML pipeline in the presence of poor-quality input images. For an exemplary ML pipeline, our new method increases the accuracy on an unseen dataset from 0.79 to 0.87.
Highlights
With the advent of high-resolution whole-slide imaging technology and the advances in deep learning, computer-aided diagnosis (CAD) systems have become a very important part of the clinical work today
1The content of sections 3.1 and 3.2 of this journal article is partially based on results from our previously published workshop paper [14]. This journal paper includes new material as follows: (a) the optimized stain normalization system presented in Stanisavljevic et al [14] is enhanced with automatic detection of the background pixels threshold; (b) we report the run-time of the optimized system and the speedup gains over the baseline implementation using a larger set of 175 images; (c) we identify a new challenge of the stain normalization algorithm, namely its sensitivity to the quality of the input images, and propose a method to deal with low-quality input images; (d) we evaluate the benefits of the new method address two challenges of the stain normalization algorithm: [1] long processing time, and [2] large system memory consumption, when normalizing high-resolution images
We continue the section with depicting a method that detects such images and normalizes them using an algorithm variant which is robust to over the original normalization algorithm by employing machine learning; (e) we show that our new normalization method increases the classification accuracy of our Machine learning (ML) pipeline when compared with the original normalization algorithm
Summary
With the advent of high-resolution whole-slide imaging technology and the advances in deep learning, computer-aided diagnosis (CAD) systems have become a very important part of the clinical work today. Machine learning (ML) based image analysis algorithms applied to digitized histological slides can assist the pathologists in terms of workload reduction, efficient decision support, and interpretability of the results [1,2,3]. One of the main pre-processing algorithms in whole-slide image (WSI) analysis is the color normalization of stained tissue samples [4]. Despite the standardized staining protocols, variations in the staining results are still frequent due to differences in, e.g., the antigen concentration and incubation time and temperature, the different conditions across slide scanners etc. Stain normalization methods aim to help the CAD systems by generating images with a standardized appearance of the different stains [5,6,7,8,9,10,11,12]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
