Deep Learning-based Image Segmentation Research Articles

Assessment of Deep Learning Based Image Segmentation for Identifying Floating Net Cages from Very High-Resolution Capella Synthetic Aperture Radar (SAR) Data

Assessment of Deep Learning Based Image Segmentation for Identifying Floating Net Cages from Very High-Resolution Capella Synthetic Aperture Radar (SAR) Data

Deep learning-derived splenic radiomics, genomics, and coronary artery disease.

Despite advances in managing traditional risk factors, coronary artery disease (CAD) remains the leading cause of mortality. Circulating hematopoietic cells influence risk for CAD, but the role of a key regulating organ, spleen, is unknown. The understudied spleen is a 3-dimensional structure of the hematopoietic system optimally suited for unbiased radiologic investigations toward novel mechanistic insights. Deep learning-based image segmentation and radiomics techniques were utilized to extract splenic radiomic features from abdominal MRIs of 42,059 UK Biobank participants. Regression analysis was used to identify splenic radiomics features associated with CAD. Genome-wide association analyses were applied to identify loci associated with these radiomics features. Overlap between loci associated with CAD and the splenic radiomics features was explored to understand the underlying genetic mechanisms of the role of the spleen in CAD. We extracted 107 splenic radiomics features from abdominal MRIs, and of these, 10 features were associated with CAD. Genome-wide association analysis of CAD-associated features identified 219 loci, including 35 previously reported CAD loci, 7 of which were not associated with conventional CAD risk factors. Notably, variants at 9p21 were associated with splenic features such as run length non-uniformity. Our study, combining deep learning with genomics, presents a new framework to uncover the splenic axis of CAD. Notably, our study provides evidence for the underlying genetic connection between the spleen as a candidate causal tissue-type and CAD with insight into the mechanisms of 9p21, whose mechanism is still elusive despite its initial discovery in 2007. More broadly, our study provides a unique application of deep learning radiomics to non-invasively find associations between imaging, genetics, and clinical outcomes.

Semantic segmentation and thermal imaging for forest fires detection and monitoring by drones

Forest ecosystems play a crucial role in providing a wide range of ecological, social, and economic benefits. However, the increasing frequency and severity of forest fires pose a significant threat to the sustainability of forests and their functions, highlighting the need for early detection and swift action to mitigate damage. The combination of drones and artificial intelligence, particularly deep learning, proves to be a cost-effective solution for accurately and efficiently detecting forest fires in real-time. Deep learning-based image segmentation models can not only be employed for forest fire detection but also play a vital role in damage assessment and support reforestation efforts. Furthermore, the integration of thermal cameras on drones can significantly enhance the sensitivity in forest fire detection. This study undertakes an in-depth analysis of recent advancements in deep learning-based semantic segmentation, with a particular focus on model’s mask region convolutional neural network (Mask R-CNN) and you only look once (YOLO) v5, v7, and v8 variants. Emphasis is placed on their suitability for forest fire monitoring using drones equipped with RGB and/or thermal cameras. The conducted experiments have yielded encouraging outcomes across various metrics, underscoring its significance as an invaluable asset for both fire detection and continuous monitoring endeavors.

Proposing an Efficient Deep Learning Algorithm Based on Segment Anything Model for Detection and Tracking of Vehicles through Uncalibrated Urban Traffic Surveillance Cameras

In this study, we present a novel approach leveraging the segment anything model (SAM) for the efficient detection and tracking of vehicles in urban traffic surveillance systems by utilizing uncalibrated low-resolution highway cameras. This research addresses the critical need for accurate vehicle monitoring in intelligent transportation systems (ITS) and smart city infrastructure. Traditional methods often struggle with the variability and complexity of urban environments, leading to suboptimal performance. Our approach harnesses the power of SAM, an advanced deep learning-based image segmentation algorithm, to significantly enhance the detection accuracy and tracking robustness. Through extensive testing and evaluation on two datasets of 511 highway cameras from Quebec, Canada and NVIDIA AI City Challenge Track 1, our algorithm achieved exceptional performance metrics including a precision of 89.68%, a recall of 97.87%, and an F1-score of 93.60%. These results represent a substantial improvement over existing state-of-the-art methods such as the YOLO version 8 algorithm, single shot detector (SSD), region-based convolutional neural network (RCNN). This advancement not only highlights the potential of SAM in real-time vehicle detection and tracking applications, but also underscores its capability to handle the diverse and dynamic conditions of urban traffic scenes. The implementation of this technology can lead to improved traffic management, reduced congestion, and enhanced urban mobility, making it a valuable tool for modern smart cities. The outcomes of this research pave the way for future advancements in remote sensing and photogrammetry, particularly in the realm of urban traffic surveillance and management.

Deep learning-based image segmentation for instantaneous flame front extraction

This paper delves into the methodology employed in examining lean premixed turbulent flame fronts extracted from Planar Laser Induced Fluorescence (PLIF) images at elevated pressures. In such flow regimes, the PLIF signal suffers from significant collisional quenching, typically resulting in image data with low signal-to-noise ratio (SNR). This poses severe difficulties for conventional flame front extraction algorithms based on intensity gradients and requires intense user intervention to yield acceptable results. In this work, we propose Convolutional Neural Network (CNN)-based Deep Learning (DL) models as an alternative to problem specific conventional methods. The pretrained DL models were fine-tuned, employing data augmentation, on a small annotated dataset including a variety of conditions between SNR ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 1.6 to 2.6 and subsequently evaluated. All DL models significantly outperformed the best-scoring conventional implementation both quantitatively and visually, while having similar inference times. IoU-scores and Recall values were found to be up to a factor ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 1.2 and ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 2.5 higher, respectively, with ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 1.15 times improved Precision. Small-scale structures were captured much better with fewer erroneous predictions, becoming particularly pronounced for the lower SNR data investigated. Moreover, by applying artificially modeled noise, it was shown that the range of image conditions in terms of SNR that can be reliably processed extends well beyond the images included in the training data, and satisfactory segmentation performances were found for SNR as low as ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document} 1.1. The presented DL-based flame front detection algorithm marks a methodology with significantly increased detection performance, while a similar computational effort for inference is achieved and the need for user-based parameter tuning is eliminated. It enables a very accurate extraction of instantaneous flame fronts in large image datasets where supervised processing is infeasible, unlocking unprecedented possibilities for the study of flame dynamics and instability mechanisms at industry-relevant conditions.

JOINEDTrans: Prior guided multi-task transformer for joint optic disc/cup segmentation and fovea detection

Deep learning-based image segmentation and detection models have largely improved the efficiency of analyzing retinal landmarks such as optic disc (OD), optic cup (OC), and fovea. However, factors including ophthalmic disease-related lesions and low image quality issues may severely complicate automatic OD/OC segmentation and fovea detection. Most existing works treat the identification of each landmark as a single task, and take into account no prior information. To address these issues, we propose a prior guided multi-task transformer framework for joint OD/OC segmentation and fovea detection, named JOINEDTrans. JOINEDTrans effectively combines various spatial features of the fundus images, relieving the structural distortions induced by lesions and other imaging issues. It contains a segmentation branch and a detection branch. To be noted, we employ an encoder with prior-learning in a vessel segmentation task to effectively exploit the positional relationship among vessel, OD/OC, and fovea, successfully incorporating spatial prior into the proposed JOINEDTrans framework. There are a coarse stage and a fine stage in JOINEDTrans. In the coarse stage, OD/OC coarse segmentation and fovea heatmap localization are obtained through a joint segmentation and detection module. In the fine stage, we crop regions of interest for subsequent refinement and use predictions obtained in the coarse stage to provide additional information for better performance and faster convergence. Experimental results demonstrate that JOINEDTrans outperforms existing state-of-the-art methods on the publicly available GAMMA, REFUGE, and PALM fundus image datasets. We make our code available at https://github.com/HuaqingHe/JOINEDTrans.

Evaluating the Margins of Breast Cancer Tumors by Using Digital Breast Tomosynthesis with Deep Learning: A Preliminary Assessment.

The assessment information of tumor margins is extremely important for the success of the breast cancer surgery and whether the patient undergoes a second operation. However, conducting surgical margin assessments is a time-consuming task that requires pathology-related skills and equipment, and often cannot be provided in a timely manner. To address this challenge, digital breast tomosynthesis technology was utilized to generate detailed cross-sectional images of the breast tissue and integrate deep learning algorithms for image segmentation, achieving an assessment of tumor margins during surgery. this study utilized post-operative tissue samples from 46 patients who underwent breast-conserving treatment, and generated image sets using digital breast tomosynthesis for the training and evaluation of deep learning models. Deep learning algorithms effectively identifying the tumor area. They achieved a Mean Intersection over Union (MIoU) of 0.91, global accuracy of 99%, weighted IoU of 44%, precision of 98%, recall of 83%, F1 score of 89%, and dice coefficient of 93% on the training dataset; for the testing dataset, MIoU was at 83%, global accuracy at 97%, weighted IoU at 38%, precision at 87%, recall rate at 69%, F1 score at 76%, dice coefficient at 86%. The initial evaluation suggests that the deep learning-based image segmentation method is highly accurate in measuring breast tumor margins. This helps provide information related to tumor margins during surgery, and by using different datasets, this research method can also be applied to the surgical margin assessment of various types of tumors.

UAV imagery coupled deep learning approach for the development of an adaptive in-house web-based application for yield estimation in citrus orchard

Orchard yield estimation enables a farmer to make informed decisions. The limitations of visual inspection-based yield estimation approaches can be effectively addressed by the intervention of unmanned aerial vehicles (UAVs) and advanced image processing using deep learning algorithms. This study proposes a methodology combining deep learning-driven UAV imagery and an in-house web-based application, “DeepYield”; to measure yield in a citrus fruit orchard. The state-of-the-art deep learning object detection models SSD, Faster RCNN, YOLOv4, YOLOv5 and YOLOv7 were evaluated for detecting “harvest-ready” and “unripe” citrus fruits from the tree images. Fruit size estimation was carried out using traditional as well as deep learning-based image segmentation models. YOLOv7 outperformed other models with a mAP, Precision, Recall, and F1-Score of 86.48, 88.54, 83.66 and 86.03%, respectively. The developed solution was integrated into a web-based application as ‘DeepYield’ to enhance users’ convenience and equip them with an automated yield estimation solution.

Spleen volume is independently associated with non-alcoholic fatty liver disease, liver volume and liver fibrosis

Non-alcoholic fatty liver disease (NAFLD) can lead to irreversible liver damage manifesting in systemic effects (e.g., elevated portal vein pressure and splenomegaly) with increased risk of deadly outcomes. However, the association of spleen volume with NAFLD and related type 2-diabetes (T2D) is not fully understood. The UK Biobank contains comprehensive health-data of 500,000 participants, including clinical data and MR images of >40,000 individuals. The present study estimated the spleen volume of 37,066 participants through automated deep learning-based image segmentation of neck-to-knee MR images. The aim was to investigate the associations of spleen volume with NAFLD, T2D and liver fibrosis, while adjusting for natural confounders. The recent redefinition and new designation of NAFLD to metabolic dysfunction-associated steatotic liver disease (MASLD), promoted by major organisations of studies on liver disease, was not employed as introduced after the conduct of this study. The results showed that spleen volume decreased with age, correlated positively with body size and was smaller in females compared to males. Larger spleens were observed in subjects with NAFLD and T2D compared to controls. Spleen volume was also positively and independently associated with liver fat fraction, liver volume and the fibrosis-4 score, with notable volumetric increases already at low liver fat fractions and volumes, but not independently associated with T2D. These results suggest a link between spleen volume and NAFLD already at an early stage of the disease, potentially due to initial rise in portal vein pressure.

Natalizumab reduces loss of gray matter and thalamic volume in patients with relapsing-remitting multiple sclerosis: A post hoc analysis from the randomized, placebo-controlled AFFIRM trial

Background: Loss of brain gray matter fractional volume predicts multiple sclerosis (MS) progression and is associated with worsening physical and cognitive symptoms. Within deep gray matter, thalamic damage is evident in early stages of MS and correlates with physical and cognitive impairment. Natalizumab is a highly effective treatment that reduces disease progression and the number of inflammatory lesions in patients with relapsing-remitting MS (RRMS). Objective: To evaluate the effect of natalizumab on gray matter and thalamic atrophy. Methods: A combination of deep learning–based image segmentation and data augmentation was applied to MRI data from the AFFIRM trial. Results: This post hoc analysis identified a reduction of 64.3% (p = 0.0044) and 64.3% (p = 0.0030) in mean percentage gray matter volume loss from baseline at treatment years 1 and 2, respectively, in patients treated with natalizumab versus placebo. The reduction in thalamic fraction volume loss from baseline with natalizumab versus placebo was 57.0% at year 2 (p < 0.0001) and 41.2% at year 1 (p = 0.0147). Similar findings resulted from analyses of absolute gray matter and thalamic fraction volume loss. Conclusion: These analyses represent the first placebo-controlled evidence supporting a role for natalizumab treatment in mitigating gray matter and thalamic fraction atrophy among patients with RRMS. ClinicalTrials.gov identifier: NCT00027300 URL: https://clinicaltrials.gov/ct2/show/NCT00027300

Automatic Detection of Collapsed Buildings after the 6 February 2023 Türkiye Earthquakes Using Post-Disaster Satellite Images with Deep Learning-Based Semantic Segmentation Models

This study focuses on the identification of collapsed buildings in satellite images after earthquakes through deep learning-based image segmentation models. The performance of four different architectures, namely U-Net, LinkNet, FPN, and PSPNet, was evaluated using various performance metrics, such as accuracy, precision, recall, F1 score, specificity, AUC, and IoU. The study used satellite images taken from the area located in the south and southeast of Türkiye covering the eleven provinces which are most affected by the Mw 7.7 Pazarcık (Kahramanmaraş) and Mw 7.6 Elbistan (Kahramanmaraş) earthquakes. The results indicated that FPN and U-Net were the best-performing models depending on the performance metric of interest. FPN achieved the highest accuracy and specificity scores, as well as the best precision score, while U-Net achieved the best recall and F1 score values, as well as the best AUC and IoU scores. The training and validation accuracy and loss curves were analyzed, and the results indicated that all four models achieved an accuracy value of over 96%. The FPN model outperformed the others in terms of accurately segmenting images while maintaining a low loss value. This study provides insights into the potential of deep learning-based image segmentation models in disaster management and can be useful for future research in this field.

A systematic review of deep learning based image segmentation to detect polyp

A systematic review of deep learning based image segmentation to detect polyp

Displacement Measurement and 3D Reconstruction of Segmental Retaining Wall Using Deep Convolutional Neural Networks and Binocular Stereovision

Computer vision techniques were employed to monitor the displacement of retaining walls using artificial markers, traditional feature detection algorithms, and photogrammetry‐based point cloud reconstruction. However, the use of artificial markers often increases both installation time and costs, whereas the performance of traditional feature matching is affected by uneven illumination, and photogrammetry techniques require multiple images for point cloud reconstruction. To overcome these limitations, a nontarget‐based displacement monitoring method for segmental retaining walls (SRWs) using a combination of deep learning and stereovision was proposed. Binocular stereovision was employed to reconstruct the geometry and surface properties of the SRW in a digital three‐dimensional (3D) model. Deep learning models were then used to extract natural features from SRW blocks, enabling displacement calculation without using artificial targets. The performance was evaluated by monitoring the behaviors of SRW experiments at both laboratory and field scales. The deep learning–based image segmentation models identified SRW block features in the experiment and real case datasets with an average F1 score from 0.910 to 0.965 under various environmental conditions. The reconstructed results of point cloud coordinates demonstrated high accuracy, ranging from 95.2% to 98.6%. Furthermore, the calculated displacement exhibited a high degree of agreement with the measured displacement. The accuracy of the calculated displacements for the laboratory and field experiments ranged from 89.5% to 99.1%. The proposed method can be used for automatic SRW displacement monitoring.

Insights into the hydrogen evolution reaction in vanadium redox flow batteries: A synchrotron radiation based X-ray imaging study

The parasitic hydrogen evolution reaction (HER) in the negative half-cell of vanadium redox flow batteries (VRFBs) causes severe efficiency losses. Thus, a deeper understanding of this process and the accompanying bubble formation is crucial. This benchmarking study locally analyzes the bubble distribution in thick, porous electrodes for the first time using deep learning-based image segmentation of synchrotron X-ray micro-tomograms. Each large three-dimensional data set was processed precisely in less than one minute while minimizing human errors and pointing out areas of increased HER activity in VRFBs. The study systematically varies the electrode potential and material, concluding that more negative electrode potentials of −200 mV vs. reversible hydrogen electrode (RHE) and lower cause more substantial bubble formation, resulting in bubble fractions of around 15%–20% in carbon felt electrodes. Contrarily, the bubble fractions stay only around 2% in an electrode combining carbon felt and carbon paper. The detected areas with high HER activity, such as the border subregion with more than 30% bubble fraction in carbon felt electrodes, the cutting edges, and preferential spots in the electrode bulk, are potential-independent and suggest that larger electrodes with a higher bulk-to-border ratio might reduce HER-related performance losses. The described combination of electrochemical measurements, local X-ray micro-tomography, AI-based segmentation, and 3D morphometric analysis is a powerful and novel approach for local bubble analysis in three-dimensional porous electrodes, providing an essential toolkit for a broad community working on bubble-generating electrochemical systems.

Deep Learning-Based Automatic Defect Detection of Additive Manufactured Stainless Steel

Accumulating interest from academia and industry, the part of quality assurance in metal additive manufacturing (AM) is achieving incremental recognition owing to its distinct advantages over conventional manufacturing methods. In this paper, we introduced a convolutional neural network, YOLOv8 approach toward robust metallographic image quality inspection. Metallographic images accommodate key information relating to metal properties, such as structural strength, ductility, toughness, and defects, which are employed to select suitable materials for multiple engineering execution. Therefore, by comprehending the microstructures, one can understand insights into the behavior of a metal component and make predictive assessments of failure under specific conditions. Deep learning-based image segmentation is a robust technique for the detection of microstructural defects like cracks, inclusion, and gas porosity. Therefore, we improvise the YOLOv8 with dilated convolution mechanisms to acquire automatic micro-structure defect characterization. More specifically, for the first time, the YOLOv8 algorithm was proposed in the metallography dataset from additive manufacturing of steels (Metal DAM) to identify defects like cracks and porosity as a novel approach. A total of 414 images from ArcelorMittal engineers were used as an open-access database. The experimental results demonstrated that the YOLOv8 model successfully detected and identified cracks and porosity in the metal AM dataset, achieving an improved defect detection accuracy of up to 96% within just 0.5 h compared to previous automatic defect recognition processes.

Application of deep learning in medical imaging segmentation

The increasing demand for segmentation of lesions in medical images necessitates research on automatic segmentation. Manual segmentation is inefficient due to training time and energy constraints. Deep learning-based image segmentation technology can improve efficiency and aid in diagnosing conditions. This technology provides accurate and detailed data support for clinical medicine, making it a crucial field in medical image processing. This essay introduces image segmentation and its classification, and explains the differences between two segmentation methods, semantic segmentation and instance segmentation and their respective application fields. Additionally, it introduces several well-known deep neural networks for segmenting medical images using deep learning. Regarding the models, this article introduces the structure and characteristics of each model as well as their respective advantages and disadvantages. This essay also introduces examples of deep learning using different deep neural network models to segment specific medical images, including image segmentation based on FCN for the heart and U-Net for the kidneys.

Segmentation of lung lobes and lesions in chest CT for the classification of COVID-19 severity

To precisely determine the severity of COVID-19-related pneumonia, computed tomography (CT) is an imaging modality beneficial for patient monitoring and therapy planning. Thus, we aimed to develop a deep learning-based image segmentation model to automatically assess lung lesions related to COVID-19 infection and calculate the total severity score (TSS). The entire dataset consisted of 124 COVID-19 patients acquired from Chulabhorn Hospital, divided into 28 cases without lung lesions and 96 cases with lung lesions categorized severity by radiologists regarding TSS. The model used a 3D-UNet along with DenseNet and ResNet models that had already been trained to separate the lobes of the lungs and figure out the percentage of lung involvement due to COVID-19 infection. It also used the Dice similarity coefficient (DSC) to measure TSS. Our final model, consisting of 3D-UNet integrated with DenseNet169, achieved segmentation of lung lobes and lesions with the Dice similarity coefficients of 91.52% and 76.89%, respectively. The calculated TSS values were similar to those evaluated by radiologists, with an R2 of 0.842. The correlation between the ground-truth TSS and model prediction was greater than that of the radiologist, which was 0.890 and 0.709, respectively.

Comparison of deep learning-based image segmentation methods for intravascular ultrasound on retrospective and large image cohort study

ObjectivesThe aim of this study was to investigate the generalization performance of deep learning segmentation models on a large cohort intravascular ultrasound (IVUS) image dataset over the lumen and external elastic membrane (EEM), and to assess the consistency and accuracy of automated IVUS quantitative measurement parameters.MethodsA total of 11,070 IVUS images from 113 patients and pullbacks were collected and annotated by cardiologists to train and test deep learning segmentation models. A comparison of five state of the art medical image segmentation models was performed by evaluating the segmentation of the lumen and EEM. Dice similarity coefficient (DSC), intersection over union (IoU) and Hausdorff distance (HD) were calculated for the overall and for subsets of different IVUS image categories. Further, the agreement between the IVUS quantitative measurement parameters calculated by automatic segmentation and those calculated by manual segmentation was evaluated. Finally, the segmentation performance of our model was also compared with previous studies.ResultsCENet achieved the best performance in DSC (0.958 for lumen, 0.921 for EEM) and IoU (0.975 for lumen, 0.951 for EEM) among all models, while Res-UNet was the best performer in HD (0.219 for lumen, 0.178 for EEM). The mean intraclass correlation coefficient (ICC) and Bland–Altman plot demonstrated the extremely strong agreement (0.855, 95% CI 0.822–0.887) between model's automatic prediction and manual measurements.ConclusionsDeep learning models based on large cohort image datasets were capable of achieving state of the art (SOTA) results in lumen and EEM segmentation. It can be used for IVUS clinical evaluation and achieve excellent agreement with clinicians on quantitative parameter measurements.

Seaweed Growth Monitoring with a Low-Cost Vision-Based System.

In this paper, we introduce a method for automated seaweed growth monitoring by combining a low-cost RGB and stereo vision camera. While current vision-based seaweed growth monitoring techniques focus on laboratory measurements or above-ground seaweed, we investigate the feasibility of the underwater imaging of a vertical seaweed farm. We use deep learning-based image segmentation (DeeplabV3+) to determine the size of the seaweed in pixels from recorded RGB images. We convert this pixel size to meters squared by using the distance information from the stereo camera. We demonstrate the performance of our monitoring system using measurements in a seaweed farm in the River Scheldt estuary (in The Netherlands). Notwithstanding the poor visibility of the seaweed in the images, we are able to segment the seaweed with an intersection of the union (IoU) of 0.9, and we reach a repeatability of 6% and a precision of the seaweed size of 18%.

Investigation on the effect of data quality and quantity of concrete cracks on the performance of deep learning-based image segmentation

The dataset is crucial for the results of crack segmentation in deep learning. However, the quantity and quality of annotations in datasets used for crack segmentation are uneven, and there is a lack of dataset benchmarks. To address this issue, this paper proposes a dataset benchmark for pixel-level segmentation of concrete cracks based on a deep learning model. Firstly, the ability of Resnet-101, DRN, Xception, and Mobilenetv2 to extract crack features is evaluated to preferably select an appropriate backbone network for the DeepLabv3 + model. Secondly, based on the deep learning model and crack dataset, a universal calculation model is proposed to estimate the quantity benchmark of high-quality crack annotated samples, as well as the mixture ratio of high-quality and low-quality annotated samples. Additionally, a pixel-level visualization characterization method is constructed for inference results of crack edge details. The results show that the model with Resnet-101 as the backbone network exhibits significant performance, with pixel accuracy (PA), mean intersection-over-union (mIoU), and frequency-weighted intersection-over-union (FWIoU) reaching 99.76%, 95.05%, and 99.52%, respectively. The combination ratio of low-quality and high-quality annotated crack images has different effects on the model performance for datasets with different number of samples. The proposed benchmark for the concrete crack dataset holds engineering value and significance, and can be applied to improve model performance by accurately selecting appropriate methods, such as increasing the number of samples in the dataset or optimizing the model architecture, while minimizing the annotation time of the dataset.