No abstract

No abstract

Deep neural networks have been extensively researched in the field of document image classification to improve classification performance and have shown excellent results. However, there is little research in this area that addresses the question of how well these models would perform in a real-world environment, where the data the models are confronted with often exhibits various types of noise or distortion. In this work, we present two separate benchmark datasets, namely RVL-CDIP-D and Tobacco3482-D, to evaluate the robustness of existing state-of-the-art document image classifiers to different types of data distortions that are commonly encountered in the real world. The proposed benchmarks are generated by inserting 21 different types of data distortions with varying severity levels into the well-known document datasets RVL-CDIP and Tobacco3482, respectively, which are then used to quantitatively evaluate the impact of the different distortion types on the performance of latest document image classifiers. In doing so, we show that while the higher accuracy models also exhibit relatively higher robustness, they still severely underperform on some specific distortions, with their classification accuracies dropping from ~90% to as low as ~40% in some cases. We also show that some of these high accuracy models perform even worse than the baseline AlexNet model in the presence of distortions, with the relative decline in their accuracy sometimes reaching as high as 300-450% that of AlexNet. The proposed robustness benchmarks are made available to the community and may aid future research in this area.

<div>Discrimination between cell types in the co-culture environment with multiple cell lines can assist in examining the interaction between different cell populations. Identifying different cell cultures along with segmentation in co-culture is essential for understanding the cellular mechanisms associated with disease states. Extracting the information from the co-culture models can help in quantifying the sub-population response to treatment conditions. In the past, there exists minimal progress related to cell-type aware segmentation in the monoculture and no development whatsoever for the co-culture. The introduction of the LIVECell dataset has provided us with the opportunity to perform experiments for cell-type aware segmentation. However, it is composed of microscopic images in a monoculture environment. In this paper, we have proposed a pipeline for coculture microscopic images data generation, where each image can contain multiple cell cultures. In addition, we have proposed a pipeline for culture-dependent cell segmentation in monoculture and co-culture microscopic images. Based on extensive evaluation, it was revealed that it is possible to achieve good quality cell-type aware segmentation in mono- and co-culture microscopic images.</div>