Blood cell counting based on U-Net++ and YOLOv5

Abstract

Clinical information about a variety of disorders is available through blood cell counting, which is usually done by manual methods. However, manual methods are complex, time-consuming and susceptible to the subjective experience of inspectors. Although many efforts have been made to develop automated blood cell counting algorithms, the complexity of blood cell distribution and the highly overlapping nature of some red blood cells (RBCs) remain significant challenges that limit the improvement of analytical accuracy. Here, we proposed an end-to-end method for blood cell counting based on deep learning. Firstly, U-Net++ was used to segment the whole blood cell image into several regions of interest (ROI), and each ROI contains only one single cell or multiple overlapping cells. Subsequently, YOLOv5 was used to detect blood cells in each ROI. Specifically, we proposed several strategies, including fine classification of RBCs, adaptive adjustment for non-maximal suppression (NMS) threshold and blood cell morphology constraints to improve the accuracy of detection. Finally, the detection outcomes for each ROI were combined and superimposed. The results show that our method can effectively address the issue of high overlap and precisely segment and detect blood cells, with a 98.18% accuracy rate for blood cell counting.

References

[1] ThejashwiniM, PadmaM C. Counting of RBC’s and WBC’s using image processing technique[J]. International journal on recent and innovation trends in computing and communication, 2015, 3(5):2948-2953

[2] OthmanM Z, MohammedT S, AliA B. Neural network classification of white blood cell using microscopic images[J]. International journal of advanced computer science and applications, 2017, 8(5):99-104

[3] KhanS, KhanA, KhattakF S, et al.. An accurate and cost effective approach to blood cell count[J]. International journal of computer applications, 2012, 50(1):975-8887

[4] MazalanS M, MahmoodN H, RazakM A A. Automated red blood cells counting in peripheral blood smear image using circular Hough transform[C], 2013, New York, IEEE: 320-324

[5] AlomariY M, Sheikh AbdullahSNH, Zaharatul AzmaR, et al.. Automatic detection and quantification of WBCs and RBCs using iterative structured circle detection algorithm[J]. Computational and mathematical methods in medicine, 2014, 2014(8):979302

[6] LiQ, ZhouM, LiuH, et al.. Red blood cell count automation using microscopic hyperspectral imaging technology[J]. Applied spectroscopy, 2015, 69(12): 1372-1380

[7] AparnaV, SarathT V, RamachandranK I. Simulation model for anemia detection using RBC counting algorithms and watershed transform[C], 2017, New York, IEEE: 284-291

[8] LecunY, BengioY, HintonG. Deep learning[J]. Nature, 2015, 521(7553):436-444

[9] GuJ, WangZ, KuenJ, et al.. Recent advances in convolutional neural networks[J]. Pattern recognition, 2018, 77: 354-377

[10] ChowdhuryA B, RobersonJ, HukkooA, et al.. Automated complete blood cell count and malaria pathogen detection using convolution neural network[J]. IEEE robotics and automation letters, 2020, 5(2):1047-1054

[11] ChourasiyaS, RaniG U. Automatic red blood cell counting using watershed segmentation[J]. Hemoglobin, 2014, 14: 17

[12] DhiebN, GhazzaiH, BesbesH, et al.. An automated blood cells counting and classification framework using mask R-CNN deep learning model[C], 2019, New York, IEEE: 300-303

[13] WangG, ZhaoT, FangZ, et al.. Experimental evaluation of deep learning method in reticulocyte enumeration in peripheral blood[J]. International journal of laboratory hematology, 2021, 43(4):597-601

[14] ZhangD, ZhangP, WangL. Cell counting algorithm based on YOLOv3 and image density estimation[C], 2019, New York, IEEE: 920-924

[15] InchurV B, PraveenL S, ShankpalP. Implementation of blood cell counting algorithm using digital image processing techniques[C], 2020, New York, IEEE: 21-26

[16] JiangN, YuF. A cell counting framework based on random forest and density map[J]. Applied sciences, 2020, 10(23):8346

[17] TessemaA W, MohammedM A, SimegnG L, et al.. Quantitative analysis of blood cells from microscopic images using convolutional neural network[J]. Medical & biological engineering & computing, 2021, 59(1): 143-152

[18] ZhouY, WangY, WuJ, et al.. ErythroidCounter: an automatic pipeline for erythroid cell detection, identification and counting based on deep learning[J]. Multimedia tools and applications, 2022, 81(18):25541-25556

[19] NARDO-MARINO A, BRAUNSTEIN T H, PETERSEN J, et al. Automating pitted red blood cell counts using deep neural network analysis: a new method for measuring splenic function in sickle cell anaemia[J]. Frontiers in physiology, 2022, 13.

[20] ZhouZ, Rahman SiddiqueeM M, TajbakhshN, et al.. Unet++: a nested U-net architecture for medical image segmentation[J]. Lecture notes in computer science volume, 2018, 11045(5):13-21

[21] JOCHER G, STOKEN A, BOROVEC J, et al. Yolov5: version 5.0[EB/OL]. (2021-04-12) [2022-05-20]. https://github.com/ultralytics/yolov5.

[22] RonnebergerO, FischerP, BroxT. U-net: convolutional networks for biomedical image segmentation[C], 2015, Heidelberg, Springer International Publishing: 234-241

[23] LinT Y, DollarP, GirshickR, et al.. Feature pyramid networks for object detection[C], 2017, New York, IEEE: 2117-2125

[24] RenS, HeK, GirshickR, et al.. Faster R-CNN towards real-time object detection with region proposal networks[J]. IEEE transactions on pattern analysis and machine intelligence, 2015, 39(6):1137-1149

[25] GirshickR. Fast R-CNN[C], 2015, New York, IEEE: 1440-1448