Visual Counting and Automated Image-analytic Assessment of Ki-67 and their Prognostic Value in Synovial Sarcoma

Background: Ki-67 is a widely used proliferation marker reflecting prognosis in various tumors. However, visual assessment and scoring of Ki-67 suffers from marked inter-observer and intra-observer variability. We aimed to assess the concordance of manual counting and automated image-analytic scoring methods for Ki-67 in synovial sarcoma. Patients and Methods: Tissue microarrays from 34 patients with synovial sarcoma were immunostained for Ki-67 and scored both visually and with 3DHistech QuantCenter. Results: The automated assessment of Ki-67 expression was in good agreement with the visually counted Ki-67 (rPearson=0.96, p<0.001). In a Cox regression model automated [hazard ratio (HR)=1.047, p=0.024], but not visual (HR=1.063, p=0.053) assessment method associated high Ki-67 scores with worse overall survival. Conclusion: The automated Ki-67 assessment method appears to be comparable to the visual method in synovial sarcoma and had a significant association to overall survival.

Label-free Viability Assay using Holographic Video Microscopy

Total Holographic Characterization (THC) is presented here as an efficient, automated, label-free method of accurately identifying cell viability. THC is a single-particle characterization technology that determines the size and index of refraction of individual particles using the Lorenz-Mie theory of light scattering. Although assessment of cell viability is a challenge in many applications, including biologics manufacturing, traditional approaches often include unreliable labeling with dyes and/or time consuming methods of manually counting cells. In this work we measured the viability of Saccharomyces cerevisiae yeast in the presence of various concentrations of isopropanol as a function of time. All THC measurements were performed in the native environment of the sample with no dilution or addition of labels. We compared our results with THC to manual counting of living and dead cells as distinguished with trypan blue dye. Our findings demonstrate that THC can effectively distinguish living and dead yeast cells by the index of refraction of individual cells.

Multi-Input Convolutional Neural Networks for Automatic Pollen Classification

Pollen allergies are a cause of much suffering for an increasing number of individuals. Current pollen monitoring techniques are lacking due to their reliance on manual counting and classification of pollen by human technicians. In this study, we present a neural network architecture capable of distinguishing pollen species using data from an automated particle measurement device. This work presents an improvement over the current state of the art in the task of automated pollen classification, using fluorescence spectrum data of aerosol particles. We obtained a relative reduction in the error rate of over 48%, from 27% to 14%, for one of the datasets, with similar improvements for the other analyzed datasets. We also use a novel approach for doing hyperparameter tuning for multiple input networks.

Selective Hole Filling of Red Blood Cells for Improved Marker-Controlled Watershed Segmentation

Manual counting and evaluation of red blood cells with the presence of malaria parasites is a tiresome, time-consuming process that can be altered by environmental conditions and human error. Many algorithms were presented to segment red blood cells for subsequent parasitemia evaluation by machine learning algorithms. However, the segmentation of overlapping red blood cells always has been a challenge. Marker-controlled watershed segmentation is one of the methods that was implemented to separate overlapping red blood cells. However, a high number of overlapped red blood cells were still an issue. We propose a novel approach to improve the segmentation efficiency of marker-controlled watershed segmentation. Local minimum histogram background segmentation with a selective hole filling algorithm was introduced to improve segmentation efficiency of marker-controlled watershed segmentation on a high number of overlapping red blood cells. The local minimum was selected on the smoothed histogram for background segmentation. The combination of selective filling, convex hull, and Hough circle detection algorithms was utilized for the intact segmentation of red blood cells. The markers were computed from the resulted mask, and finally, marker-controlled watershed segmentation was applied to separate overlapping red blood cells. As a result, the proposed algorithm achieved higher background segmentation accuracy compared to popular background segmentation algorithms, and the inclusion of corner details improved watershed segmentation efficiency.

A Rapid Detection Method for Tomato Gray Mold Spores in Greenhouse Based on Microfluidic Chip Enrichment and Lens-Less Diffraction Image Processing

It is of great significance to find tomato gray mold in time and take corresponding control measures to ensure the production of tomato crops. This study proposed a rapid detection method for spores of Botrytis cinerea in green-house based on microfluidic chip enrichment and lens-free diffraction image processing. Microfluidic chip with a regular triangular inner rib structure was designed to achieve the enrichment of Botrytis cinerea spores. In order to obtain the diffraction image of the diseased spores, a lens-less diffraction imaging system was built. Furthermore, the collected spore diffraction images were processed and counted. The simulation results showed that the collection efficiency of 16 μm particles was 79%, 100%, and 89% at the inlet flow rate of 12, 14 and 16 mL/min, respectively. The experimental verification results were observed under a microscope. The results showed that when the flow rate of the microfluidic chip was 12, 14 and 16 mL/min, the collection efficiency of Botrytis cinerea spores was 70.65%, 87.52% and 77.96%, respectively. The Botrytis cinerea spores collected in the experiment were placed under a microscope for manual counting and compared with the automatic counting results based on diffraction image processing. A total of 10 sets of experiments were carried out, with an error range of the experiment was 5.13~8.57%, and the average error of the experiment was 6.42%. The Bland–Altman method was used to analyze two methods based on diffraction image processing and manual counting under a microscope. All points are within the 95% consistency interval. Therefore, this study can provide a basis for the research on the real-time monitoring technology of tomato gray mold spores in the greenhouse.

Counting using deep learning regression gives value to ecological surveys

Many ecological studies rely on count data and involve manual counting of objects of interest, which is time-consuming and especially disadvantageous when time in the field or lab is limited. However, an increasing number of works uses digital imagery, which opens opportunities to automatise counting tasks. In this study, we use machine learning to automate counting objects of interest without the need to label individual objects. By leveraging already existing image-level annotations, this approach can also give value to historical data that were collected and annotated over longer time series (typical for many ecological studies), without the aim of deep learning applications. We demonstrate deep learning regression on two fundamentally different counting tasks: (i) daily growth rings from microscopic images of fish otolith (i.e., hearing stone) and (ii) hauled out seals from highly variable aerial imagery. In the otolith images, our deep learning-based regressor yields an RMSE of 3.40 day-rings and an $$R^2$$ R 2 of 0.92. Initial performance in the seal images is lower (RMSE of 23.46 seals and $$R^2$$ R 2 of 0.72), which can be attributed to a lack of images with a high number of seals in the initial training set, compared to the test set. We then show how to improve performance substantially (RMSE of 19.03 seals and $$R^2$$ R 2 of 0.77) by carefully selecting and relabelling just 100 additional training images based on initial model prediction discrepancy. The regression-based approach used here returns accurate counts ($$R^2$$ R 2 of 0.92 and 0.77 for the rings and seals, respectively), directly usable in ecological research.

Comparison of different autoanalyzers for the determination of lymphocyte and neutrophil counts in mouse blood

Objectives Autoanalyzers are used in clinical haematology for analysis of blood samples in clinical as well as in nonclinical studies. The results from these analyzers vary from machine to machine. In this study, we compared the lymphocyte and neutrophil count of mouse blood between ADVIA 2120i, Horiba Yumizen H2500 and CellaVision analyzers against manual counting as gold standard. Methods Blood samples from 28 female BALB/c mice were collected and analyzed. Agreement between different autoanalyzers and manual counting were determined by Bland–Altman method. Results A high level of agreement was found between CellaVision and manual technique for lymphocyte (Bias=4.75, 95% limits of agreement −14 to 24) and neutrophil count (Bias=0.68 (−17 to 19)). Agreement in lymphocyte count was also observed between ADVIA and manual counting, but to a lesser extent compared to CellaVision (Bias=13.9 (−10.45 to 38.27)). However, no agreement was observed for ADVIA (Neutrophils), Horiba (lymphocytes and neutrophils) with manual counting. Conclusions Our data suggests that CellaVision could be used for the differential counting of neutrophil and lymphocytes in mouse blood sample.

Machine Learning-Based Plant Detection Algorithms to Automate Counting Tasks Using 3D Canopy Scans

This study tested whether machine learning (ML) methods can effectively separate individual plants from complex 3D canopy laser scans as a prerequisite to analyzing particular plant features. For this, we scanned mung bean and chickpea crops with PlantEye (R) laser scanners. Firstly, we segmented the crop canopies from the background in 3D space using the Region Growing Segmentation algorithm. Then, Convolutional Neural Network (CNN) based ML algorithms were fine-tuned for plant counting. Application of the CNN-based (Convolutional Neural Network) processing architecture was possible only after we reduced the dimensionality of the data to 2D. This allowed for the identification of individual plants and their counting with an accuracy of 93.18% and 92.87% for mung bean and chickpea plants, respectively. These steps were connected to the phenotyping pipeline, which can now replace manual counting operations that are inefficient, costly, and error-prone. The use of CNN in this study was innovatively solved with dimensionality reduction, addition of height information as color, and consequent application of a 2D CNN-based approach. We found there to be a wide gap in the use of ML on 3D information. This gap will have to be addressed, especially for more complex plant feature extractions, which we intend to implement through further research.

Label-free Viability Assay using Holographic Video Microscopy

Total Holographic Characterization (THC) is presented here as an efficient, automated, label-free method of accurately identifying cell viability. THC is a single-particle characterization technology that determines the size and index of refraction of individual particles using the Lorenz-Mie theory of light scattering. Although assessment of cell viability is a challenge in many applications, including biologics manufacturing, traditional approaches often include unreliable labeling with dyes and/or time consuming methods of manually counting cells. In this work we measured the viability of Saccharomyces cerevisiae yeast in the presence of various concentrations of isopropanol as a function of time. All THC measurements were performed in the native environment of the sample with no dilution or addition of labels. We compared our results with THC to manual counting of living and dead cells as distinguished with trypan blue dye. Our findings demonstrate that THC can effectively distinguish living and dead yeast cells by the index of refraction of individual cells.

Automatic Detection and Counting of Blood Cells in Smear Images Using RetinaNet

A complete blood count is one of the significant clinical tests that evaluates overall human health and provides relevant information for disease diagnosis. The conventional strategies of blood cell counting include manual counting as well as counting using the hemocytometer and are tedious and time-consuming tasks. This research-based paper proposes an automatic software-based alternative method to count blood cells accurately using the RetinaNet deep learning network, which is used to recognize and classify objects in microscopic images. After training, the network automatically recognizes and counts red blood cells, white blood cells, and platelets. We tested a model trained on smear images and found that the trained model has generalized capabilities. We assessed the quality of detection and cell counting using performance measures, such as accuracy, sensitivity, precision, and F1-score. Moreover, we studied the dependence of the confidence thresholds and the number of learning epochs on the obtained results of recognition and counting. We compared the performance of the proposed approach with those obtained by other authors who dealt with the subject of cell counting and show that object detection and labeling can be an additional advantage in the task of counting objects.