Histopathological Image Deep Feature Representation for CBIR in Smart PACS

Abstract

Pathological Anatomy is moving toward computerizing processes mainly due to the extensive digitization of histology slides that resulted in the availability of many Whole Slide Images (WSIs). Their use is essential, especially in cancer diagnosis and research, and raises the pressing need for increasingly influential information archiving and retrieval systems. Picture Archiving and Communication Systems (PACSs) represent an actual possibility to archive and organize this growing amount of data. The design and implementation of a robust and accurate methodology for querying them in the pathology domain using a novel approach are mandatory. In particular, the Content-Based Image Retrieval (CBIR) methodology can be involved in the PACSs using a query-by-example task. In this context, one of many crucial points of CBIR concerns the representation of images as feature vectors, and the accuracy of retrieval mainly depends on feature extraction. Thus, our study explored different representations of WSI patches by features extracted from pre-trained Convolution Neural Networks (CNNs). In order to perform a helpful comparison, we evaluated features extracted from different layers of state-of-the-art CNNs using different dimensionality reduction techniques. Furthermore, we provided a qualitative analysis of obtained results. The evaluation showed encouraging results for our proposed framework. Graphical Abstract

Creating an Atlas of Normal Tissue for Pruning WSI Patching Through Anomaly Detection

“…While most of the computational pathology tasks are designed to classify or detect the presence of pathological lesions in gigapixel WSIs, the presence of computationally redundant normal tissue in WSIs is often ignored. While a large body of scientific work has been done in the field of WSI retrieval systems to address the numerical representation of the WSI 12,[23][24][25] and WSI search engines 7,8,26,27 , no study has addressed the effect of normal tissue in the indexing and search pipeline. There- clusion as part of their workflow 28 , the potential effect of redundant normal tissue in WSI retrieval is an unknown area.…”

Benchmarking Deep Learning-based Image Retrieval of Oral Tumor Histology

“…Therefore, content-based image retrieval (CBIR) is suitable. CBIR regards histopathological images as query images to find similar images from a database based on their similar morphology [2,5]. This system is useful as a diagnostic aid for finding case references, especially where diagnostic expertise is challenging to find, such as in low-to middle-income countries [1].…”

“…The CBIR system consists of two aspects: image feature extraction and nearest-neighbor search. Feature extraction is crucial because it must adequately capture complex histological features such as staining patterns, tissue structures, and cellular morphology to create histologically relevant image representation [2,7,8]. The extracted features must be robust to irrelevant color variations, such as different hematoxylin and eosin (HE) stain brands, glass slide color degradation, and image-capturing devices ranging from whole-slide image (WSI) scanners to smartphone cameras [5,8,9].…”

Patient Re-Identification Based on Deep Metric Learning in Trunk Computed Tomography Images Acquired from Devices from Different Vendors

“…Image retrieval is a computer vision task that aims to find images similar to an image query from a large-scale image database such as PACS [ 9 – 11 ]. Recently, as an applied technology of content-based image retrieval to solve patient misidentification, image analysis techniques that link the correct patient information with an image, called patient re-identification, have shown promising results for X-ray images [ 2 , 4 , 12 , 13 – 20 ], two-dimensional scout computed tomography (CT) [ 21 , 22 ], three-dimensional (3D) scout magnetic resonance images [ 23 ], and 3D CT [ 24 ] images.…”