Template Matching Research Articles

Synergistically segmenting and reducing fracture bones via whole-to-whole deep dense matching

In bone fracture preoperative surgical planning, 3D fracture reduction is of great significance. However, segmenting 3D fracture bones into fragments and assembling the bone fragments to restore their original morphology are labor intensive and time consuming. Current works, however, omit their synergistic effects between reduction and segmentation, which, we envision, can significantly benefit the fracture reduction. To this end, a method is proposed to alternatively segment and assemble the fragments by incorporating the contralateral bone as template and leveraging the synergistic effect on segmentation and reduction. The 3D fracture model is initially segmented into separated outer surface of fragments. The reduction is then conducted by template matching, specifically, a whole-to-whole matching strategy through deep learning of dense features is employed. Thereafter, the segmentation is further refined with the assistance of the template. The process of reduction and segmentation is iterated until the algorithm converges. Experiments were conducted on simulated data and clinical data, the mean segmentation accuracy is 95.26%, the assembling translation error is 3.28 ± 3.01 mm, and the assembling rotation error is 1.62 ± 1.71°. Results demonstrated the superiority of our method over state-of-the-art methods, and it can be used as a feasible scheme for fracture reduction planning. Moreover, it is also proved that segmentation and reduction can promote each other.

Neuromorphic computing with hybrid CNN–Stochastic Reservoir for time series WiFi based human activity recognition

Wi-Fi Channel State Information (CSI) based human activity recognition (HAR) which using channel disturbances caused by signal reflection is a novel way of environment sensing and motion recognition. The collected channels characteristics are heavily influenced by the environment, human activity patterns and subject’s weight and height. These signal variations reflected from body components are mainly affected by static multipath effects comprises random noise and behave differently in individuals, and thus an active field of research. To reach further for achieving automated real-time classification, lower computational cost and easy adaptability to hardware are necessary. In this work, a CSI-based HAR with hybrid framework, Convolutional Neural Network (CNN)-Stochastic Reservoir (SR) (CNN-SR) has been proposed, enabling a subject adaptable and more efficient hardware implementation with minimal computational complexity. A subcarrier correlation matrix is first computed and portrayed in image without preprocessing based on the reflection of the raw CSI signal induced by human activities at regular intervals, allowing visual observation of whole pattern changes. The time-based features are subsequently extracted through CNN and these feature arrays are then feed into SR which based on stochastic spiking neural network (SSNN) in simple cycle reservoir architecture for template matching. SR offers attractive power savings over typical von Neumann systems, by doing stochastic computations. The proposed method has also been demonstrated that is capable for HAR based on partially captured signals. The signal pattern of each segment can be observed in a single sight and then employed for person-to-person template recognition. This enables HAR with minimal computational complexity and solving the inter-person variability concerns. The results demonstrate that the proposed CNN-SR achieves impressive performance in recognizing human activities and surpasses existing models with an average accuracy of 93.49%.

Pose estimation for swimmers in video surveillance

Traditional models for pose estimation in video surveillance are based on graph structures, in this paper, we propose a method that breaks the limitation of template matching within a range of pose changes to obtain robust results. We implement our swimmer pose estimation method based on deep learning. We take use of High-Resolution Net (HRNet) to extract and fuse visual features of visual object and complete the object detection using the key points of human joint. The proposed model could be applied to all kinds of swimming styles throughout appropriate training. Compared with the methods that require multimodel combinations and training, the proposed method directly achieves the end-to-end prediction, which is easily to be implemented and deployed. In addition, a cross-fusion module is added between parallel networks, which assists the network to make use of the characteristics of multiple resolutions. The proposed network has achieved ideal results in the pose estimation of swimmers by comparing HRNet-W32 and HRNet-W48. In addition, we propose an annotated key point dataset of swimmers which was created from the view of underwater swimmers. Compared with side view, the torso of swimmers collected by the underwater view is much suitable for a broad spectrum of machine vision tasks.

Bionic visual navigation model for enhanced template matching and loop closing in challenging lighting environments

In recent years, various computational models have been proposed in neuroscience, but only a few have been utilized in robot visual navigation through engineering approaches. Besides, these engineering methods lack environmental adaptability, especially in the aspect of visual place recognition in weak or uncontrollable light fluctuations. To address this issue and enhance the performance of visual template matching and map loop closure detection in challenging lighting environments, this paper proposed a bionic visual navigation model that combines two neural network architectures, the Pulse Coupled Neural Network (PCNN) and Continuous Attractor Neural Network (CANN). In our navigation model, the visual features of the environments are encoded as temporal information by the spiking model and input into the local view cells for visual template matching. Additionally, we utilize the pose cells network, which incorporates the similarity between current and previous templates, along with odometry data, to encode spatial information and store it as an experience map. To validate the effectiveness of the proposed model, we conducted evaluations on datasets collected within our library, campus, and an open-source office dataset. The experimental results reveal that our algorithm increases the F1-score of template matching by approximately 10.5%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ \\% $$\\end{document}, 35.8%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ \\% $$\\end{document}, 61.7%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ \\% $$\\end{document}, and 1.9%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$ \\% $$\\end{document} in each dataset, compared to the conventional RatSLAM method. Furthermore, our algorithm generates a more accurate map that closely correlates with the real-world trajectory without compromising on computation time. The results suggest that our bionic visual navigation model is reliable for both standard and extreme lighting conditions.

Research on 3D image reconstruction of sparse power lines by array GM-APD lidar

Power lines and other tiny obstacles in the air are deadly hazards to safe aircraft flight. However, existing detection methods have difficulty sensing power lines’ weak radiation signals submerged in noise, making power lines detection an international research hotspot. To solve this problem, this paper proposes the use of array Geiger-mode avalanche photodiode (GM-APD) light detection and ranging (lidar) with single photon detection sensitivity to achieve remote three-dimensional (3D) imaging detection of power lines (GM-PLTR). We focus on solving the problem of noise filtering for sparse power lines under low Signal-to-noise ratio (SNR) conditions and completing 3D image reconstruction. Firstly, we propose a method based on adaptive thresholding and waveform density discrimination combined with spatial dependency to remove the pixel-wise noise, including environmental light, noise echoes from neighboring pixels, and dark count noise. Then we obtain the high SNR depth image of the targets. Secondly, we use an improved DBSCAN algorithm to separate the power lines from the background noise in the depth image to improve the reconstruction effect. Finally, we use a one-directional template matching method to fill in the missing shapes of the power lines. Experimental results show that the algorithm improves the SNR of the power lines echo signal by an average of 2.3 times and achieves sparse power lines imaging detection at 340.95 m. Moreover, the Hough transform confirms the possibility of high-precision detection of power lines, significantly reducing the difficulty of detecting small targets. Finally, it will help unmanned aerial vehicles (UAVs) complete the inspection of power lines.

Extensive Angular Sampling Enables the Sensitive Localization of Macromolecules in Electron Tomograms.

Cryo-electron tomography provides 3D images of macromolecules in their cellular context. To detect macromolecules in tomograms, template matching (TM) is often used, which uses 3D models that are often reliable for substantial parts of the macromolecules. However, the extent of rotational searches in particle detection has not been investigated due to computational limitations. Here, we provide a GPU implementation of TM as part of the PyTOM software package, which drastically speeds up the orientational search and allows for sampling beyond the Crowther criterion within a feasible timeframe. We quantify the improvements in sensitivity and false-discovery rate for the examples of ribosome identification and detection. Sampling at the Crowther criterion, which was effectively impossible with CPU implementations due to the extensive computation times, allows for automated extraction with high sensitivity. Consequently, we also show that an extensive angular sample renders 3D TM sensitive to the local alignment of tilt series and damage induced by focused ion beam milling. With this new release of PyTOM, we focused on integration with other software packages that support more refined subtomogram-averaging workflows. The automated classification of ribosomes by TM with appropriate angular sampling on locally corrected tomograms has a sufficiently low false-discovery rate, allowing for it to be directly used for high-resolution averaging and adequate sensitivity to reveal polysome organization.

Spine Stereotactic Body Radiation Therapy Without Immobilization: Detailed Analysis of Intrafraction Motion Using High-Frequency kV Imaging During Irradiation

Spine SBRT requires high positioning accuracy and a stable patient to maximize target coverage and reduce excessive irradiation to organs at risk (OARs). Positional verification during spine SBRT delivery helps to ensure accurate positioning for all patients. We report our experience with non-invasive 3-dimensional (3D) target position monitoring during volumetric modulated arc therapy (VMAT) of spine metastases, in non-immobilized patients positioned using only a thin mattress and simple arm and knee supports. Fluoroscopic planar kV images were acquired at 7 frames/second using the on-board imaging system during VMAT spine SBRT. Template matching and triangulation were used to track the target in vertical, longitudinal and lateral direction. If the tracking trace deviated >1mm from the planned position in ≥1 direction, treatment was manually interrupted and 6D CBCT-based couch correction was performed. Tracking data was used to retrospectively analyze the target position. Positional data, agreement with CBCT, correlation between position of the couch and direction of any positional correction, and treatment times were analyzed. 175 fractions were analyzed. Delivery was interrupted 83 times in 66 fractions for a deviation >1mm. In 97% of cases the difference between tracking data and subsequent clinical shift performed after the CBCT match was ≤0.5mm. Lateral/longitudinal shift performed after intervention correlated with the couch roll/pitch at the start of treatment (correlation coefficient: -0.63/0.53). Mean (SD; range) time between start of first imaging and end of the last arc was 15.2 minutes (5.1;7.6-36.3). Spine tracking during irradiation can be used to prompt an intervention CBCT scan and repositioning, so that a spine SBRT target deviates by ≤1mm from the planned position, even in non-immobilized patients. kV tracking and CBCT are in good agreement. The data support verification CBCT after all 6 degrees-of-freedom positional corrections in non-immobilized spine SBRT patients.

Towards robust registration of heterogeneous multispectral UAV imagery: A two-stage approach for cotton leaf lesion grading

Multiple source images acquired from diverse sensors mounted on unmanned aerial vehicles (UAVs) offer valuable complementary information for ground vegetation analysis. However, accurately aligning heterogeneous UAV images poses challenges due to differences in geometry, intensity, and noise resulting from varying imaging principles. This paper presents a two-stage registration method aimed at fusing visible RGB and multispectral images for cotton leaf lesion grading. The coarse alignment stage utilizes Scale Invariant Feature Transform (SIFT), while the refined alignment stage employs a novel correlation coefficient-based template matching. The proposed method first employs the EfficientDet network to detect infected cotton leaves with lesions in RGB images. Subsequently, lesion leaves in multiple spectral imagery (red, green, red edge, and near-infrared bands) are located using the perspective transformation matrix derived from SIFT and the coordinates of lesion leaves in RGB images. Refined registration between RGB and multispectral imagery is achieved through template matching with the new correlation coefficient. The registered reflectance data from the different spectral bands and RGB components are utilized to classify pixels in each infected leaf into lesion, healthy, and soil parts. The lesion grade is determined based on the ratio of lesion pixels to the total corresponding leaf area. Experimental results, compared with manual assessment, demonstrate a lesion leaves detection model with a mAP@0.5 of 91.01% and a leaf lesion grading accuracy of 92.01%. These results validate the suitability of the proposed method for UAV RGB and multispectral image registration, enabling automated cotton leaf lesion grading.

Security System Enhancement using Iris Scan Biometric Technology for ATM Machine

Banking is currently very simple, however there are often many opportunities for cybercrime. Numerous banking transactions have been the victim of fraud. Concerns around consumer security and privacy have always revolved around authentication and verification. It is challenging to uphold the integrity and authenticity of people in the ever-shifting environment. On how to combat ATM fraud, many ideas and recommendations have recently been put out. We need some trustworthy security findings that we can utilize in conjunction with the current technologies to combat all of these frauds. One of the technologies that we can use in conjunction with the existing technology is biometrics. Combining related technology may help to decrease ATM fraud and hence increase the security of other financial transactions. This essay offers an overview of important iris recognition studies. An iris identification system goes through three main phases: template matching, point birth, and picture preprocessing. The most popular algorithms used in each stage are reviewed in the literature.

Seismology in the cloud: guidance for the individual researcher

The commercial cloud offers on-demand computational resources that could be revolutionary for the seismological community, especially as seismic datasets continue to grow. However, there are few educational examples for cloud use that target individual seismological researchers. Here, we present a reproducible earthquake detection and association workflow that runs on Microsoft Azure. The Python-based workflow runs on continuous time-series data using both template matching and machine learning. We provide tutorials for constructing cloud resources (both storage and computing) through a desktop portal and deploying the code both locally and remotely on the cloud resources. We report on scaling of compute times and costs to show that CPU-only processing is generally inexpensive, and is faster and simpler than using GPUs. When the workflow is applied to one year of continuous data from a mid-ocean ridge, the resulting earthquake catalogs suggest that template matching and machine learning are complementary methods whose relative performance is dependent on site-specific tectonic characteristics. Overall, we find that the commercial cloud presents a steep learning curve but is cost-effective. This report is intended as an informative starting point for any researcher considering migrating their own processing to the commercial cloud.

Automated cataloging of oyster toadfish (Opsanus tau) boatwhistle calls using template matching and machine learning

Oyster toadfish (Opsanus tau) represent an ecologically significant species found throughout estuaries along the eastern coast of the United States. While these crevice-dwelling fish can be challenging to observe in their habitats, it is possible to infer their distribution and aspects of their behavior by recording the sounds they produce. The task of cataloging the distinctive advertisement boatwhistle sounds produced by male toadfish to attract females throughout the spring and summer is automated using a multi-step process. Candidate boatwhistles are first identified by template matching using a suite of synthetic spectrogram kernels formed to mimic the two lowest frequency harmonic tones within the boatwhistle. Candidate boatwhistle calls are identified based on the correlation between these kernels and a low-frequency spectrogram of the data. Next, frequency-reassigned spectrogram images of these candidates are formed and input into the pre-trained ResNet-50 convolutional neural network. Finally, the activations from a deep, fully connected layer within this network are extracted and passed to a one-vs-all support-vector-machine classifier, which separates boatwhistles from the larger set of candidate signals. This classifier model was trained and evaluated using a labeled dataset of over 20,000 candidate signals generated over diverse acoustic conditions within Pamlico Sound, North Carolina, USA. The accompanying software provides an effective and efficient tool to monitor boatwhistle calls, which may facilitate a deeper understanding of the spatial distribution, behavioral patterns, and ecological roles played by oyster toadfish.

Detection and Positioning of Camellia oleifera Fruit Based on LBP Image Texture Matching and Binocular Stereo Vision

To achieve the rapid recognition and accurate picking of Camellia oleifera fruits, a binocular vision system composed of two industrial cameras was used to collect images of Camellia oleifera fruits in natural environments. The YOLOv7 convolutional neural network model was used for iterative training, and the optimal weight model was selected to recognize the images and obtain the anchor frame region of the Camellia oleifera fruits. The local binary pattern (LBP) maps of the anchor frame region were extracted and matched by using the normalized correlation coefficient template matching algorithm to obtain the positions of the center point in the left and right images. The recognition experimental results showed that the accuracy rate, recall rate, mAP and F1 of the model were 97.3%, 97.6%, 97.7% and 97.4%. The recognition rate of the Camellia oleifera fruit with slight shading was 93.13%, and the recognition rate with severe shading was 75.21%. The recognition rate of the Camellia oleifera fruit was 90.64% under sunlight condition, and the recognition rate was 91.34% under shading condition. The orchard experiment results showed that, in the depth range of 400–600 mm, the maximum error value of the binocular stereo vision system in the depth direction was 4.279 mm, and the standard deviation was 1.142 mm. The detection and three-dimensional positioning accuracy of the binocular stereo vision system for Camellia oleifera fruits could basically meet the working requirements of the Camellia oleifera fruit-picking robot.

False positives are common in single-station template matching

Template matching has become a cornerstone technique of observational seismology. By taking known events, and scanning them against a continuous record, new events smaller than the signal-to-noise ratio can be found, substantially improving the magnitude of completeness of earthquake catalogues. Template matching is normally used in an array setting, however as we move into the era of planetary seismology, we are likely to apply template matching for very small arrays or even single stations. Given the high impact of planetary seismology studies on our understanding of the structure and dynamics of non-Earth bodies, it is important to assess the reliability of template matching in the small-n setting. Towards this goal, we estimate a lower bound on the rate of false positives for single-station template matching by examining the behaviour of correlations of totally uncorrelated white noise. We find that, for typical processing regimes and match thresholds, false positives are likely quite common. We must therefore be exceptionally careful when considering the output of template matching in the small-n setting.

A fast matching method for skip displacement recognition of hoisting system

Determining if there is any coal left in the coal mine skip unloading is a challenging problem. We propose a new object tracking method to provide new ideas to solve the problem of whether there is residual coal in the skip unloading. In this study, a fast vision-based measurement framework about new diamond search normalized cross-correlation (NDS-NCC), including a movement prediction part and a template tracking part, is used. Firstly, a diamond search method with fast adaptive coarse localization is proposed to overcome the time-consuming problem of NCC template matching process. Secondly, we incorporate a template updating strategies that allows targets to be tracked consistently. The performance of the NDS-NCC algorithm was tested on the bench and at the coal mine site for balancing cylinders and wire rope tracking. The results show that the NDS-NCC algorithm has better efficiency (7 ms/frame) and accuracy (0.4512 px) in the coal mine field application.

An Extended Labanotation Generation Method Based on 3D Human Pose Estimation for Intangible Cultural Heritage Dance Videos

To address the issues of low accuracy in existing 3D human pose estimation (HPE) methods and the limited level of details in Labanotation, we propose an extended Labanotation generation method for intangible cultural heritage dance videos based on 3D HPE. First, a 2D human pose sequence of the performer is inputted along with spatial location embeddings, where multiple spatial transformer modules are employed to extract spatial features of human joints and generate cross-joint multiple hypotheses. Afterward, temporal features are extracted by a self-attentive module and the correlation between different hypotheses is learned using bilinear pooling. Finally, the 3D joint coordinates of the performer are predicted, which are matched with the corresponding extended Labanotation symbols using the Laban template matching method to generate extended Labanotation. Experimental results show that, compared with VideoPose and CrossFormer algorithms, the Mean Per Joint Position Error (MPJPE) of the proposed method is reduced by 3.7[Formula: see text]mm and 0.6[Formula: see text]mm, respectively on Human3.6M dataset, and the generated extended Labanotation can better describe the movement details compared with the basic Labanotation.

VERIFIKASI DOKUMEN TRANSKRIP NILAI SEMESTER MENGGUNAKAN METODA OPTICAL CHARACTER RECOGNITION

At the Ambon State Polytechnic, students' semester grade reports are still manually typed. This causes frequent typo errors which can result in the invalidity of the document, let alone incorrect grades, student identification numbers and many other label values. Here a java application has been implemented to detect these errors. This application is primarily intended for officials of the Head of Study Program, Head of the Department before signing and validating the report. Officials who legalize it will be greatly assisted because tedious validation work can be replaced by computers. The validation process is carried out by utilizing the optical character recognition technique from the open source library Tesseract-OCR. From the experimental results the verification process can be improved by using OCR specific on specific regions of interest (ROI) after using template matching method from OpenCV. The consideration of the Levehnstein distance in the comparison of label values against the reference database also improves the success rate of the algorithm. The database used has been tested for about 800 grade report documents, with successful verification result above 90%.

Application of Enhanced Optimal-Detection of Time–Frequency Domain Reflectometry on HTS Cable With High-Resolution

Time–frequency domain reflectometry (TFDR) has been used for sensing various types of cables including high temperature superconducting (HTS) cables. Recently, the optimal-detection of time–frequency domain reflectometry (OD-TFDR) has been proposed and has better detection and localization performance than conventional TFDR. Additionally, OD-TFDR has provided a threshold that can distinguish detected signals from Gaussian noise. However, OD-TFDR does not provide a criterion for classifying the detected signals into fault signals and artifacts. Furthermore, the diagnosis result of OD-TFDR is not intuitive because an additional tool is needed to locate each fault. Thus, this paper proposes a new diagnosis algorithm that can provides a criterion for distinguishing between artifacts and faults by considering the propagation direction of the incident signal in the join time–frequency domain while preserving the detection and localization performance of OD-TFDR. Through a template matching process of generalized time–frequency cross-correlation (GTFCC), the core algorithm of OD-TFDR, a modified–generalized time–frequency cross-correlation (M–GTFCC) function is derived. Next, a modified–time–frequency cross-correlation (M–TFCC) function is derived by slicing the M–GTFCC function considering the propagation direction of signal. Finally, M–TFCC function is proposed as a new intuitive diagnosis result with artifact-free and high-resolution. The performance of the proposed algorithm is tested via experimental setup for a real-world single phase HTS cable with an emulated local quench by heating element, and the efficacy of the proposed algorithm is verified based on three performances: fault detection, locating accuracy, and resolution.

Hole Detection in Plastic Mulch Using Template Matching and Machine Learning Algorithms

Mulch is a ground cover material to maintain soil moisture and temperature stability as a plant medium. Mulch also helps prevent weed growth for better plant growth. For planting with plastic mulch, farmers need to make holes in the mulch the day before planting. Precision agriculture is needed because it can obtain savings in input financing, labor, and better yields, so this research aims to identify holes in mulch based on Unmanned Aerial Vehicle images. The advantage of this research is that it can monitor each plant based on the mulch holes, and the number of holes identified can be used as a parameter to estimate the amount of crop production. This research combines Template Matching Algorithm and Machine Learning Algorithm to improve accuracy in predicting holes in mulch. Three machine learning algorithms are used, namely the Random Forest, Support Vector Machine, and XGBoost. The data used is an orthophoto mosaic from aerial photographs. Nine areas were taken from orthophotos to be used as research samples. The results of this study obtained the highest average recall, precision, and f-measure values using the Support Vector Machine algorithm with a recall value of 87.7%, precision of 97.5%, and f-score of 92.3%. This research focuses on reducing detected commission errors. Therefore, omission errors were still detected in the damaged or leaf-covered holes.

A new microarchitecture hardware proposal for the use of the improved template matching method in face similarity detection

Template matching is a technique used in many different areas in the literature to search for a small template image in a larger source image. The execution time of the technique can be quite high depending on the size of the source image, since the search made in the large image is realized pixel by pixel in the form of frames selected in the template size. This situation creates inconveniences in terms of real-time use of the technique. In addition, the results cannot be interpreted correctly due to the fact that there are many fake matches in the region where there are real matches. In this study, the technique was strengthened with a new and simple proposal called neighborhood pool to prevent fake matches. In order to reduce the long scan time to real-time execution time, the reinforced template matching technique was redesigned as an FPGA-based hardware architecture. This new micro-hardware architecture, which includes the neighborhood pool idea, was synthesized and embedded in the FPGA device by time analysis and tested on the device. For the tests of the study, the face matching scenario was chosen and the face image of a particular person was searched in the larger source image. The improved template matching technique in software-based tests reduced the execution time by about 4.5 times compared to the classical template matching method, and about 300 times in hardware testing, and it was observed that the system execution time became compatible with real-time. When the proposed new micro hardware architecture is compared with the CNN and template matching designs in the literature, it is clearly seen that it has a very low hardware requirement compared to them.

Context-guided coarse-to-fine detection model for bird nest detection on high-speed railway catenary

As a critical component of ensuring the safe and stable operation of trains, the detection of bird’s nests on the rail catenary has always been essential. Low-resolution images and the lack of labelled data, however, make it difficult to detect smaller bird’s nests (those occupying small pixels in the input image). Previous solution relies on manual online patrol or offline video playback, which severely limits the detection efficiency. Previously, this challenge was addressed by manual online patrol or offline video playback, which severely limits detection efficiency. We propose in this work a context-guided coarse-to-fine detection model (CG-CFDM) for solving the bird’s nest detection problem. This solution consists of a context reasoning module and a coarse-to-fine detection network. By detecting domains and matching templates, the context reasoning module generates new labelled context bounding boxes, thereby reducing the burden of annotation. As a result of its delicately designed architecture and powerful representation learning ability, this trained coarse-to-fine detection network further facilitates the detection of bird’s nests in an efficient and accurate manner. Extensive experiments demonstrate that the proposed approach is superior to existing methods in terms of performance and has a great deal of potential for detecting bird’s nests.