Objective and Impact Statement: Human epidermal growth factor receptor 2 (HER2) is a critical protein in cancer cell growth that signifies the aggressiveness of breast cancer (BC) and helps predict its prognosis. Here, we introduce a deep learning-based approach utilizing pyramid sampling for the automated classification of HER2 status in immunohistochemically (IHC) stained BC tissue images. Introduction: Accurate assessment of IHC-stained tissue slides for HER2 expression levels is essential for both treatment guidance and understanding of cancer mechanisms. Nevertheless, the traditional workflow of manual examination by board-certified pathologists encounters challenges, including inter- and intra-observer inconsistency and extended turnaround times. Methods: Our deep learning-based method analyzes morphological features at various spatial scales, efficiently managing the computational load and facilitating a detailed examination of cellular and larger-scale tissue-level details. Results: This approach addresses the tissue heterogeneity of HER2 expression by providing a comprehensive view, leading to a blind testing classification accuracy of 84.70%, on a dataset of 523 core images from tissue microarrays. Conclusion: This automated system, proving reliable as an adjunct pathology tool, has the potential to enhance diagnostic precision and evaluation speed, and might substantially impact cancer treatment planning.

Systemic amyloidosis involves the deposition of misfolded proteins in organs/tissues, leading to progressive organ dysfunction and failure. Congo red is the gold-standard chemical stain for visualizing amyloid deposits in tissue, showing birefringence under polarization microscopy. However, Congo red staining is tedious and costly to perform, and prone to false diagnoses due to variations in amyloid amount, staining quality and manual examination of tissue under a polarization microscope. We report virtual birefringence imaging and virtual Congo red staining of label-free human tissue to show that a single neural network can transform autofluorescence images of label-free tissue into brightfield and polarized microscopy images, matching their histochemically stained versions. Blind testing with quantitative metrics and pathologist evaluations on cardiac tissue showed that our virtually stained polarization and brightfield images highlight amyloid patterns in a consistent manner, mitigating challenges due to variations in chemical staining quality and manual imaging processes in the clinical workflow.

Fabrication and characterization of flexible optical fiber bundles (FBs) with in-house synthesized high-index and low-index thermally matched glasses are presented. The FBs composed of around 15000 single-core fibers with pixel sizes between 1.1 and 10 μm are fabricated using the stack-and-draw technique from sets of thermally matched zirconium-silicate ZR3, borosilicate SK222, sodium-silicate K209, and F2 glasses. With high refractive index contrast pair of glasses ZR3/SK222 and K209/F2, FBs with numerical apertures (NAs) of 0.53 and 0.59 are obtained, respectively. Among the studied glass materials, ZR3, SK222, and K209 are in-house synthesized, while F2 is commercially acquired. Seven different FBs with varying pixel sizes and bundle diameters are characterized. Brightfield imaging of a micro-ruler and a Convallaria majalis sample and fluorescence imaging of a dye-stained paper tissue and a cirrhotic mice liver tissue are demonstrated using these FBs, demonstrating their good potential for microendoscopic imaging. Brightfield and fluorescence imaging performance of the studied FBs are compared. For both sets of glass compositions, good imaging performance is observed for FBs, with core diameter and core-to-core distance values larger than 1.6 μm and 2.3 μm, respectively. FBs fabricated with K209/F2 glass pairs revealed better performance in fluorescence imaging due to their higher NA of 0.59.

Data sorting is used in many fields and plays an important role in defining the overall speed and performance. There are - many sorting categories. In this study, two of these sorting algorithms that are bitonic sort and radix sort are dealt with. We have designed and developed Radix Sort and Bitonic Sort algorithms for many core Graphics Processing Units (GPUs). Although bitonic sort is a concurrent sorting algorithm, radix sort is a distribution sorting algorithm, i.e. both of these algorithms are not usual sorting algorithms. They can be parallelized on GPUs easily to get better performance than other sorting algorithms. We parallelized these sorting algorithms on many core GPUs using the Compute Unified Device Architecture (CUDA) platform, developed by NVIDIA Corporation and got some performance measurements.

Code embedding represents code semantics in vector form. Although code embedding-based systems have been successfully applied to various source code analysis tasks, further research is required to enhance code embedding for better code analysis capabilities, aiming to surpass the performance and functionality of static code analysis tools. In addition, standard methods for improving code embedding are essential to develop more effective embedding-based systems, similar to augmentation techniques in the image processing domain. This study aims to create a contrastive learning-based system to explore the potential of a generic method for enhancing code embedding for code classification tasks. A triplet loss-based deep learning network is designed to optimize in-class similarity and increase the distance between classes. An experimental dataset that contains code from Java, Python, and PHP programming languages and 4 different code smells is created by collecting code from open-source repositories on GitHub. We evaluate the proposed system’s effectiveness with widely used BERT, CodeBERT, and GraphCodeBERT pretrained models to create code embedding for the code classification task of code smell detection. Our findings indicate that the proposed system may offer improvements in accuracy, an average of 8% and a maximum of 13% for models. These results suggest that incorporating contrastive learning techniques into the generation process of code representation as a preprocessing step can enhance performance in code analysis.