Blood Cells Counting and Localisation through Deep Learning Object Detection

Object detection is a technology that deals with recognizing classes of objects and their location. It is used in many different areas, such as in face-detecting systems [16, 34, 37], surveillance tools [9], human-machine interfaces [17], and self-driving cars [18, 23, 25, 26, 30]. These days, deep learning object detection approaches have achieved significantly better performance than the classical feature-based algorithms. Darknet [31] is a deep learning object detection framework, which is well known for its fast speed and simple structure. Unfortunately, Darknet can only work with Nvidia CUDA [6] for accelerating its deep learning calculations. For this reason, users have only limited options of selecting appropriate graphic cards. Open computing language (OpenCL) [35], an open standard for cross-platform, parallel programming of heterogeneous systems, is available for the general hardware accelerators. However, many deep learning frameworks including Darknet have no support for OpenCL. In our previous paper, we presented OpenCL-Darknet [19], which transformed the CUDA-based Darknet into an open standard OpenCL backend. The original OpenCL-Darknet successfully showed its ability for the general graphics processing unit (GPU) hardware. However, it could not achieve competitive performance compared with the CUDA version, and it only supported a limited platform. In this study, we improved the performance of OpenCL-Darknet with several optimization techniques and added support for various architectures. We also evaluated OpenCL-Darknet not only in AMD R7 accelerated processing unit (APU) with OpenCL 2.0, but also in Nvidia GPU and ARM Mali embedded GPU with OpenCL 1.2 Profile. The evaluation using the standard object detection datasets showed that our advanced OpenCL-Darknet reduced the processing time by at most 50% on average for various deep learning object detection networks compared with our original implementation. We also showed that our OpenCL deep learning framework has competitiveness compared with the CUDA-based one.

Application of Deep Learning for Object Detection

Vegetation cover classification using mixed or low-resolution scalar images is challenging. Fortunately, recently deep learning object detection methods have emerged as a replacement to the conventional machine learning methods for the detection and classification of land use and land cover. This paper presents a deep learning object detection approach for land use and land cover detection using low/mixed resolution satellite images acquired from Google Earth satellite images. Google Earth images are accessible freely using the Google Earth Pro desktop application. Our dataset consists of two (02) classes (vegetation and non-vegetation) with a total of 450 labeled images captured from different parts of Pakistan. We present a comparison of the recent anchor-free object detection model YOLOX with the anchor-based object detection model YOLOR for solving real-time problems. The end-to-end differentiability, efficient GPU utilization, and absence of hand-crafted parameters make anchor-free models a compelling choice in object detection, and yet not been explored on Land cover classification using satellite images. Our experimental study shows that YOLOX delivers an overall accuracy of 83.50% on Vegetation and 86% on Non-Vegetation classes, which outperformed YOLOR by 30% on Vegetation classes and 34% on non-Vegetation classes for our dataset. We also show how an object detection system can be used for Vegetation and Non-Vegetation classification tasks, which can then be used for change monitoring and assisting in developing geographical maps using low/mixed resolution freely available satellite images.

Background of study: Blood cell analysis is vital for diagnosing medical conditions, but traditional manual methods are laborious and error-prone. Deep learning, especially YOLO models, offers automated solutions for medical image analysis. However, the real-world effectiveness of the latest YOLOv11 in blood cell detection is not thoroughly investigated, as general object detection improvements may not translate to biomedical images due to their unique characteristics.Aims and scope of paper: This study systematically compares YOLOv10 and YOLOv11 on a public blood cell detection dataset to assess if YOLOv11's advancements provide tangible benefits for blood cell classification. The goal is to identify the most effective model for accurate and efficient detection in microscopic images, guiding AI-driven diagnostic tool selection.Methods: Both models were trained and tested under identical conditions using the Kaggle Blood Cell Detection Dataset (RBCs, WBCs, Platelets). Images were resized to 640x640 pixels. Performance metrics included mAP (mAP@50 and mAP@50–95), Precision, Recall, F1-score, speed, model complexity, and training time.Result: YOLOv11n consistently showed higher accuracy (mAP50: 0.9279 vs. 0.9120; mAP50-95: 0.6524 vs. 0.6347), particularly for RBCs and WBCs. However, YOLOv11n had longer inference (11.35 ms/image) and postprocessing times (8.64 ms/image) compared to YOLOv10n (7.00 ms/image and 0.90 ms/image). YOLOv11n trained faster (0.311 hours vs. 0.375 hours), with a smaller model size (5.5 MB vs. 5.8 MB), fewer parameters, and reduced computational complexity.Conclusion: YOLOv11n offers superior accuracy and improved training efficiency, making it suitable for medical image object detection where precision is paramount. The increased inference and postprocessing times indicate a performance-speed trade-off. Model selection should balance these factors based on deployment context.

Leucocyte classification for leukaemia detection using image processing techniques.

Air-to-ground object detection has important application value in many fields such as military reconnaissance, air strike, personnel rescue, and natural exploration. Object detection algorithm based on deep learning solves the problems of poor generality and low robustness of traditional manual detection algorithms. It has great advantages in air-to-ground detection tasks, but for small unmanned platforms, currently deep learning object detection algorithms have high hardware requirements and are difficult to deploy. In this paper, a deep learning object detection simulation system is built based on the TX2 embedded development board with the features of low power consumption, light weight and fast calculation speed. This object detection simulation experiment is performed for two mainstream deep learning object detection algorithms. The experimental results showed that the regression-based SSD algorithm combined with MobileNets feature extraction network could perform real-time in the embedded system, which laid the foundation for the actual construction of the next unmanned aerial platform.

Object counting is an indispensable task in manufacturing and management. Recently, the development of image-processing techniques and deep learning object detection has achieved excellent performance in object-counting tasks. Accordingly, we propose a novel small-size smart counting system composed of a low-cost hardware device and a cloud-based object-counting software server to implement an accurate counting function and overcome the trade-off presented by the computing power of local hardware. The cloud-based object-counting software consists of a model adapted to the object-counting task through a novel DBC-NMS (our own technique) and hyperparameter tuning of deep-learning-based object-detection methods. With the power of DBC-NMS and hyperparameter tuning, the performance of the cloud-based object-counting software is competitive over commonly used public datasets (CARPK and SKU110K) and our custom dataset of small pills. Our cloud-based object-counting software achieves an mean absolute error (MAE) of 1.03 and a root mean squared error (RMSE) of 1.20 on the Pill dataset. These results demonstrate that the proposed smart counting system accurately detects and counts densely distributed object scenes. In addition, the proposed system shows a reasonable and efficient cost–performance ratio by converging low-cost hardware and cloud-based software.

Among the blood cells, white blood cells (WBC), which play a crucial role in forming our body's defense system, are essential components. Originating in the bone marrow, these cells serve as the fundamental components of the immune system, shouldering the responsibility of safeguarding the body against foreign microbes and diseases. Insufficient WBC counts may compromise the body's skill to resist infections, a status known as leukopenia. White blood cell counting is a specialty procedure that is usually carried out by qualified physicians and radiologists. Thanks to recent advances, image processing techniques are frequently used in biological systems to identify a wide spectrum of illnesses. In this work, image processing techniques were applied to enhance the white blood cell deep learning models' classification accuracy. To expedite the classification process, Convolutional Neural Network (CNN) models were combined with Ridge feature selection and Maximal Information Coefficient techniques. These tactics successfully determined the most important characteristics. The selected feature set was then applied to the classification procedure. ResNet-50, VGG19, and our suggested model were used as feature extractors in this study. The categorizing of white blood cells was completed with an amazing 98.27% success rate. Results from the experiments demonstrated a considerable improvement in classification accuracy using the proposed CNN model.

Clinical and morphological features of a case of leukaemic reticuloendotheliosis

Leukemia is a type of cancer that is on white blood cell. This disease are characterized by abundance of abnormal white blood cell called lymphoblast in the bone marrow. Classification of blood cell types, calculation of the ratio of cell types and comparison with normal blood cells can be the subject of diagnosing this disease. The diagnostic process is carried out manually by hematologists through microscopic image. This method is likely to provide a subjective result and time-consuming.The application of digital image processing techniques and machine learning in the process of classifying white blood cells can provide more objective results. This research used thresholding method as segmentation and multilayer method of back propagation perceptron with variations in the extraction of textural features, geometry, and colors. The results of segmentation testing in this study amounted to 68.70%. Whereas the classification test shows that the combination of feature extraction of GLCM features, geometry features, and color features gives the best results. This test produces an accuration value 91.43%, precision value of 50.63%, sensitivity 56.67%, F1Score 51.95%, and specitifity 94.16%.

Adaptive Neuro Fuzzy Inference System for White Blood Cell Classification

In this paper, we demonstrate a machine learning algorithm with multiple GPU processing in hematocytes detection. Object detection in compound microscopy images presents a specific task. Microscopy image import directly to a specimen slide under the compound microscope by an image sensor device. We propose Faster R-CNN with a customizing model applying cutting-edge object detection systems. The MicrosisDCN (Microbes Diagnosis with Deep Convolutional Neural Network) deploys Faster R-CNN in python script configures to access multiple GPUs computation with 7168 CUDA cores of dual GPUs with Linux command options: worker_replicas and num_clones equal to several graphics processing units. The datasets consist of hematocytes extracted from raw slides under a microscope. These images separate the cells of interest into three groups: red blood cell (RBCs), white blood cell (WBCs), and platelets. The training dataset consists of 80 percent of the 40,000 images. And the testing dataset consists of 20 percent of the 40,000 images. Our algorithm also provides the result of mean average precision (mAP) and enables multiple GPU training models in Tensorflow and OpenCV. The mAP is the average of average precision (AP) with intersection-over-union (IoU) in measuring the score of object detection accuracy. If the mAP score approaches 1.0, it indicates well accuracy. Our 9,000 steps valuation algorithm model by the python script estimates the mAP in 3 groups: RBCs, WBCs, and platelets as about 0.9147, 0.9664, and 0.9548. The optimum would be at nearly 12,000 steps because the algorithm estimates the mAP model as 0.96 in all types. The experiment aims to verify the neural network model using a compound microscope.

Many diseases such as anemia and leukemia are detected by counting Red Blood Cells (RBCs or erythrocytes). Generally, there are manual and automatic method for RBCs counting. In the manual method, RBCs counting is performed by a hematologist with the help of special medical equipment. The manual method is tedious, time-consuming and dependent on equipment and hematologist's expert and accuracy. The automatic method is fast, high accuracy and independent of equipment and hematologist. It is performed with the assistance of a microscopic image of the person's blood and a computer system. In this paper, a new method is presented for counting incomplete or cropped RBCs by Circular Hough Transform (CHT) and another method called EndPoints method which will be described in the following. Two parallel tasks are performed on the input image. In the first work, the input image is segmented using thresholding and erosion on the green channel. In the second work, the input image is converted to a grayscale image and edged by the Canny method. Then, the segmented image is subtracted from the edged image and remove White Blood Cells (WBCs or leucocytes) from it. After margin is added, by defining EndPoints set estimated continuation of RBCs will be added to the image. Finally, CHT is applied to the image and it will calculate the number of RBCs. Also, new counting error and counting accuracy metrics are defined for counting problem. The method is evaluated on the fifteen images of the ALL-IDB1 database and achieved 97.14 overall accuracy.

PurposeThe hematological changes in COVID-19 patients continue to receive great attention, especially in the field of public health. To our knowledge, coronavirus disease may be identified based on the severity of illness, and the study of peripheral blood smears may offer important information to facilitate the identification. Thus, we evaluated the morphological abnormalities (atypical and immature lymphocytes, lymphocytes with micronuclei, various nuclear abnormalities among erythrocytes) and quantitative changes in peripheral blood cells among 48 individuals with COVID-19 disease. MethodsThe present study is a retrospective analysis of 48 individuals, including 24 hospitalized patients diagnosed with COVID-19 disease. The blood smears of all patients were subjected to a hematological examination to identify various morphological abnormalities in white and red blood cells. In addition, a micronucleus test was conducted to assess the incidence of chromosomal damage in lymphocytes. Furthermore, the complete blood count (CBC) was performed to evaluate changes in peripheral blood cells, particularly the differential total leukocyte count, which could indicate the immune response against viral infection in COVID-19 patients. ResultsThe findings of the hematological study conducted on patients diagnosed with COVID-19 disease revealed neutrophilia, eosinophilia, mild monocytosis, decreased hematocrit level, elevated erythrocyte sedimentation rate (ESR), and immature leukocytes. It was observed that patients who were infected with coronavirus demonstrated mild thrombocytopenia. Furthermore, the micronucleus test indicated the presence of immature cells with micronuclei among lymphocytes and numerous nuclear abnormalities in red blood cells. These results help to shed light on the hematological changes that occur in COVID-19 patients, and could potentially contribute to the development of more effective treatments for the disease. ConclusionsThe examination of complete blood counts (CBCs) in conjunction with peripheral blood smears offers a potential means of identifying the impact of SARS-CoV-2 infection on the hematopoietic and immune systems, thereby providing early indications of inflammation. Based on a study, it has been suggested that SARS-CoV-2 may affect red and white blood cells causing morphological alterations thereby establishing a corresponding relationship with disease severity.

One-stage and lightweight CNN detection approach with attention: Application to WBC detection of microscopic images