Field Of View Research Articles

Measurement Approach for the Pose of Flanges in Cabin Assemblies through Distributed Vision.

The relative rotation angle between two cabins should be automatically and precisely obtained during automated assembly processes for spacecraft and aircraft. This paper introduces a method to solve this problem based on distributed vision, where two groups of cameras are employed to take images of mating features, such as dowel pins and holes, in oblique directions. Then, the relative rotation between the mating flanges of two cabins is calculated. The key point is the registration of the distributed cameras; thus, a simple and practical registration process is designed. It is assumed that there are rigid and scaling transformations among the world coordinate systems (WCS) of each camera. Therefore, the rigid-correct and scaling-correct matrices are adopted to register the cameras. An auxiliary registration device with known features is designed and moved in the cameras' field of view (FOV) to obtain the matrix parameters so that each camera acquires traces of every feature. The parameters can be solved using a genetic algorithm based on the known geometric relationships between the trajectories on the registration devices. This paper designs a prototype to verify the method. The precision reaches 0.02° in the measuring space of 340 mm.

Mathematical modeling and implementation of a panoramic zoom system without central blind area

The panoramic annular lens (PAL) is renowned for providing simultaneous vision of the panoramic space. However, its inherent central blind area is a significant drawback. In this study, we aim to harness the benefits of the PAL while mitigating its limitations by proposing a dual-channel panoramic zoom system without a central blind area. The introduction of a positive front lens group for secondary imaging is a critical aspect of our design, which is guided and substantiated by our mathematical imaging model. The results from our design affirm the practicality and effectiveness of the secondary imaging. This design can achieve a zoom ratio of 1.46× and a maximum field of view (FoV) range of 71°∼100°, enabling zooming from front to rear. Exemplifying compactness, our system measures just 60mm in total length, while providing a high-resolution output, as indicated by a modulation transfer function exceeding 0.3 at 200lp/mm. With an F−θ distortion of less than 3.3%, the system maintains low levels of distortion. The analysis is instructive and the design shows potential in situations that necessitate compliance with size and resolution requirements.

Magnetic resonance imaging of the cerebellopontine angle: comparison between constructive interference steady-state and small field-of-view technique turbo spin echo sequences.

The aim of this work was to optimize athree-dimensional (3D) turbo-spin-echo (TSE) sequence using asmall field-of-view (FOV) technique for the study of the cerebellopontine angle and to compare it with aconstructive interference steady-state (CISS) sequence. A total of 30consecutive patients underwent magnetic resonance imaging with a3Tesla (T) scanner, including 3D CISS and the optimized 3D small FOV technique turbo spin echo (3D SFT-TSE) T2-weightedsequences for the study of the cerebellopontine angle. The 3D SFT-TSE sequence was optimized after three different steps, and aquantitative evaluation of the signal-to-noise ratio (SNR) was obtained according to the National Electrical Manufacturers Association (NEMA) method. Three neuroradiologists made ablind comparative qualitative evaluation of the images between the 3D CISS and the 3D SFT-TSE obtained after the third optimization step, based on spatial resolution, contrast resolution, and presence of artifacts and noise. The calculation of SNR using the NEMA method confirmed the superiority of the third optimization step over the others. For both spatial and contrast resolution, the optimized SFT-TSE was considered better (p < 0.001) than the CISS, while image artifacts and noise were considered worse in the CISS sequence (p < 0.001). Intraobserver analysis showed that all neuroradiologists preferred the 3D SFT-TSE sequence in terms of both spatial resolution and contrast resolution and found more noise and artifact disruption in the CISS sequence. The use of the 2D radiofrequency pulse technique with a3D SFT-TSE T2 sequence was significantly more efficient than the 3D CISS sequence for the study of the cerebellopontine angle and inner ear structures.

Large Scale Turbulent Round Free Jet Far Field Investigations Using Shake-The-Box 3D Lagrangian Particle Tracking

In this study, we utilize the Shake-The-Box (STB) Lagrangian Particle Tracking (LPT) algorithm to conduct large-scale measurements in the far field of a turbulent jet. We confirm self-similar mean velocity and Reynold stress profiles across two fields of view (FOVs) spanning 90 to 240 nozzle diameters in streamwise direction. Furthermore, the utilization of a novel version of the data assimilation scheme FlowFit enables the visualization of instantaneous realizations of coherent structures in terms of Q-values. The results indicate that the structures are predominantly tube rather than hairpin-like. The high number of tracked particles and the extensive measurement range allow for the estimation of mean dissipation as a function of the distance from the nozzle using structure functions. A comparison of results derived from longitudinal and transverse approaches with those obtained using the velocity-acceleration structure function indicates that the latter provides more accurate results due to its validity in anisotropic flows. However, the findings also indicate that the anisotropy diminishes downstream. The determination of mean dissipation and acceleration enables the investigation of the validity of the Heisenberg-Yaglom relation in this anisotropic flow. Nevertheless, acceleration is a highly intermittent value, which hinders capturing all relevant events to measure its variance accurately. This paper presents the difficulties encountered in the tracking and filtering of the trajectories, offering a perspective on potential future solutions to the problem.

An Experimental Investigation Of µUAV Rotor Wingtip Vortices Through High Spatial Resolution PIV

The interest for µ UAVs platforms with VTOL capabilities dramatically rose over the last decade due to the fast-paced advancements in electronics and control. However, the vortical structures shedded from the rotary wings -and their interaction-, which are responsible for high power usage and high noise level emissions, are yet to be fully understood. In this work, an experimental PIV campaign was conducted on a commercially available propeller with a 230mm diameter. The operating angular velocity was chosen to be 5750RPM, producing a chord Reynolds number of about Re_{c,75%} ≃ 11000, generating 3.42N of thrust while requiring 55W of power. The resolution of the PIV setup was able to reliably fit 8 measurement points in the vortex core when using narrowest field-of-view (FOV), while the widest FOV comprises a measurement area of about 0.5D×0.57D. Time averaged as well as phase averaged measurements were performed. The evolution and interaction of the wingtip vortices and shear layers shedded in the flow field is captured with an azimuthal angle step increment of ∆Ψ = 5º. The thrust produced by the multirotor configuration was indeed found to be approximately 5% lower than the one produced by the isolated propeller in hovering conditions. Additionally, the PIV setup successfully met the target spatial accuracy requirements: induced velocities are accurately captured in every stage of the wingtip vortex evolution.

Pan-cortical cellular imaging in freely behaving mice using a miniaturized micro-camera array microscope (mini-MCAM).

Understanding how circuits in the brain simultaneously coordinate their activity to mediate complex ethnologically relevant behaviors requires recording neural activities from distributed populations of neurons in freely behaving animals. Current miniaturized imaging microscopes are typically limited to imaging a relatively small field of view, precluding the measurement of neural activities across multiple brain regions. Here we present a miniaturized micro-camera array microscope (mini-MCAM) that consists of four fluorescence imaging micro-cameras, each capable of capturing neural activity across a 4.5 mm x 2.55 mm field of view (FOV). Cumulatively, the mini-MCAM images over 30 mm2 area of sparsely expressed GCaMP6s neurons distributed throughout the dorsal cortex, in regions including the primary and secondary motor, somatosensory, visual, retrosplenial, and association cortices across both hemispheres. We demonstrate cortex-wide cellular resolution in vivo Calcium (Ca 2+ ) imaging using the mini-MCAM in both head-fixed and freely behaving mice.

Comparison of coronary CT angiography and invasive coronary angiography results

IntroductionCoronary artery disease (CAD) is a leading cause of death worldwide. Accurate diagnosis and management are critical. Non-invasive imaging, such as coronary computed tomography angiography (CCTA), is vital for early diagnosis and treatment planning. This study evaluates the accuracy of CAD-Reporting and Data System (CAD-RADS) scoring and the compatibility between CCTA and invasive coronary angiography (ICA) in patients suspected of having CAD.Materials and methodsFrom January 1, 2022 to January 15, 2024, 214 patients suspected of CAD underwent both CCTA and ICA. CCTA artifacts led to the exclusion of 32 patients and 128 vessels, leaving 586 vessels for analysis. CAD-RADS scoring categorized coronary stenosis. Diagnostic performance was measured by specificity, sensitivity, accuracy, positive and negative predictive value (NPV). Extracardiac findings were analyzed with a wide field of view (FOV) during CCTA.ResultsA total of 214 patients (67.3% male, median age 56) were examined. Hypertension, smoking, calcium score, and high-risk plaques correlated with CCTA and ICA CAD-RADS scores; calcium score also related to hypertension, smoking, diabetes, and dyslipidemia (p < 0.05). CCTA showed a sensitivity of 80.8% and NPV of 90.3% for detecting stenosis of 70% or more; for 50% stenosis, sensitivity was 93.5% and NPV 92.1%. Agreement between CCTA and ICA was excellent in bypass patients; stenosis detection in stented patients had 85.7% sensitivity and 96.2% NPV.ConclusionThis study highlights the importance of CAD-RADS and CCTA in CAD diagnosis and treatment planning. CCTA effectively evaluates stents and grafts, emphasizing the benefits of extracardiac findings and a wide FOV.

Design of an Airborne Low-Light Imaging System Based on Multichannel Optical Butting

For the purpose of achieving long-range, high-resolution, and ultra-wide-swath airborne earth imaging at extremely low-light levels (0.01 Lux), a low-light imaging system built on multi-detector optical butting was researched. Having decomposed the system’s specifications and verified its low-light imaging capability, we proposed to employ an optical system with a large relative aperture and low distortion and achieve imaging through the field-of-view (FOV) butting facilitated by eight 1080P high-sensitivity scientific complementary metal-oxide semiconductor (SCMOS) detectors. This paper elaborates on the design concept of the mechanical configuration of the imaging system; studies the calculation method of the structural parameters of the reflection prism; provides mathematical expressions for geometric parameters, such as the length and width of the splicing prism; and designs in detail the splicing structure of six reflection prisms for eight-channel beam splitting. Based on the design and computational results, a high-resolution, wide-swath imaging system for an ambient illuminance of 0.01 Lux was developed. Exhibiting a ground sampling distance (GSD) of 0.5 m (at a flight height of 5 km), this low-light imaging system keeps the FOV overlap ratio between adjacent detectors below 3% and boasts an effective image resolution of 4222 × 3782. The results from flight testing revealed that the proposed imaging system is capable of generating wide-swath, high-contrast resolution imagery under airborne and low-light conditions. As such, the way the system is prepared can serve as a reference point for the development of airborne low-light imaging devices.

Optical Imaging Model Based on GPU-Accelerated Monte Carlo Simulation for Deep-Sea Luminescent Objects

Modeling and simulating the underwater optical imaging process can assist in optimizing the configuration of underwater optical imaging technology. Based on the Monte Carlo (MC) method, we propose an optical imaging model which is tailored for deep-sea luminescent objects. Employing GPU parallel acceleration expedites the speed of MC simulation and ray-tracing, achieving a three-order-of-magnitude speedup over a CPU-based program. A deep-sea single-lens imaging system is constructed in the model, composed of a luminescent object, water medium, double-convex lens, aperture diaphragm, and sensor. The image of the luminescent object passing through the imaging system is generated using the forward ray-tracing method. This model enables an intuitive analysis of the inherent optical properties of water and imaging device parameters, such as sensor size, lens focal length, field of view (FOV), and camera position on imaging outcomes in the deep-sea environment.

Field-of-view extension for brain diffusion MRI via deep generative models.

In brain diffusion magnetic resonance imaging (dMRI), the volumetric and bundle analyses of whole-brain tissue microstructure and connectivity can be severely impeded by an incomplete field of view (FOV). We aim to develop a method for imputing the missing slices directly from existing dMRI scans with an incomplete FOV. We hypothesize that the imputed image with a complete FOV can improve whole-brain tractography for corrupted data with an incomplete FOV. Therefore, our approach provides a desirable alternative to discarding the valuable brain dMRI data, enabling subsequent tractography analyses that would otherwise be challenging or unattainable with corrupted data. We propose a framework based on a deep generative model that estimates the absent brain regions in dMRI scans with an incomplete FOV. The model is capable of learning both the diffusion characteristics in diffusion-weighted images (DWIs) and the anatomical features evident in the corresponding structural images for efficiently imputing missing slices of DWIs in the incomplete part of the FOV. For evaluating the imputed slices, on the Wisconsin Registry for Alzheimer's Prevention (WRAP) dataset, the proposed framework achieved , , , and ; on the National Alzheimer's Coordinating Center (NACC) dataset, it achieved , , , and . The proposed framework improved the tractography accuracy, as demonstrated by an increased average Dice score for 72 tracts ( ) on both the WRAP and NACC datasets. Results suggest that the proposed framework achieved sufficient imputation performance in brain dMRI data with an incomplete FOV for improving whole-brain tractography, thereby repairing the corrupted data. Our approach achieved more accurate whole-brain tractography results with an extended and complete FOV and reduced the uncertainty when analyzing bundles associated with Alzheimer's disease.

High-resolution, highly monochromatic, time-framed, x-ray-backlit-imaging diagnostics with dual-channel toroidal crystals

Here, we developed a high-resolution, dual-channel toroidal-crystal x-ray imager for time-framed x-ray backlit imaging diagnostics using the 4.727 keV helium-like Ti line. We also presented a method for adjusting the dual-channel imager through the self-imaging of a two-dimensional periodic object. Offline x-ray experiments achieved a spatial resolution of ∼5.0 μm in the center and better than 8.0 μm within a field of view (FOV) of ∼2 mm. At the ShenGuang-III prototype laser facility, we obtained imaging results with a spatial resolution of better than 5 μm within an FOV of ±40 µm. This imager thus provides a way of observing with high spatial, temporal, and spectral resolutions to diagnose the behavior of laser-produced plasma.

Interactions between MSTIDs and Ionospheric Irregularities in the Equatorial Region Observed on 13–14 May 2013

We investigate the interactions between medium-scale traveling ionospheric disturbances (MSTIDs) and the equatorial ionization anomaly (EIA) as well as between MSTIDs and equatorial plasma bubbles (EPBs) on the night of 13–14 May 2013, based on observations from multiple instruments (an all-sky imager, digisonde, and global positioning system (GPS)). Two dark bands (the low plasma density region) for the MSTIDs were observed moving toward each other, encountering and interacting with the EIA, and subsequently interacting again with the EIA before eventually dissipating. Then, a new dark band of MSTIDs moved in the southwest direction, drifted into the all-sky imager’s field of view (FOV), and interacted with the EIA. Following this interaction, a new dark band split off from the original dark band, slowly moved in the northeast direction, and eventually faded away in a short time. Subsequently, the original southwestward-propagating dark band of the MSTIDs encountered eastward-moving EPBs, leading to an interaction between the MSTIDs and the EPBs. Then, the dark band of the MSTIDs faded away, while the EPBs grew larger with a pronounced westward tilt. The results from various observational instruments indicate the pivotal role played by the high-density region of the EIA in the occurrence of various interaction processes. In addition, this study also revealed that MSTIDs propagating into the equatorial region can significantly impact the morphology and evolution characteristics of EPBs.

Cone-beam CT with a noncircular (sine-on-sphere) orbit: imaging performance of a clinical system for image-guided interventions.

We aim to compare the imaging performance of a cone-beam CT (CBCT) imaging system with noncircular scan protocols (sine-on-sphere) to a conventional circular orbit. A biplane C-arm system (ARTIS Icono; Siemens Healthineers) capable of circular and noncircular CBCT acquisition was used, with the latter orbit (sine-on-sphere, "Sine Spin") executing a sinusoidal motion with tilt amplitude over the half-scan orbit. A test phantom was used for the characterization of image uniformity, noise, noise-power spectrum (NPS), spatial resolution [modulation transfer function (MTF) in axial and oblique directions], and cone-beam artifacts. Findings were interpreted using an anthropomorphic head phantom with respect to pertinent tasks in skull base neurosurgery. The noncircular scan protocol exhibited several advantages associated with improved 3D sampling-evident in the NPS as filling of the null cone about the spatial frequency axis and reduction of cone-beam artifacts. The region of support at the longitudinal extrema was reduced from 16 to at a radial distance of 6.5cm. Circular and noncircular orbits exhibited nearly identical image uniformity and quantum noise, demonstrating cupping of and overall noise of . Although both the radially averaged axial MTF ( ) and 45deg oblique MTF ( ) were lower for the noncircular orbit compared with the circular orbit at the default full reconstruction field of view (FOV), there was no difference in spatial resolution for the medium reconstruction FOV (smaller voxel size). Differences in the perceptual image quality for the anthropomorphic phantom reinforced the objective, quantitative findings, including reduced beam-hardening and cone-beam artifacts about structures of interest in the skull base. Image quality differences between circular and noncircular CBCT orbits were quantitatively evaluated on a clinical system in the context of neurosurgery. The primary performance advantage for the noncircular orbit was the improved sampling and elimination of cone-beam artifacts.

Extracellular vesicle isolation and counting system (EVics) based on simultaneous tandem tangential flow filtration and large field-of-view light scattering.

Although the isolation and counting of small extracellular vesicles (sEVs) are essential steps in sEV research, an integrated method with scalability and efficiency has not been developed. Here, we present a scalable and ready-to-use extracellular vesicle (EV) isolation and counting system (EVics) that simultaneously allows isolation and counting in one system. This novel system consists of (i) EVi, a simultaneous tandem tangential flow filtration (TFF)-based EV isolation component by applying two different pore-size TFF filters, and (ii) EVc, an EV counting component using light scattering that captures a large field-of-view (FOV). EVi efficiently isolated 50-200nm-size sEVs from 15µL to 2L samples, outperforming the current state-of-the-art devices in purity and speed. EVc with a large FOV efficiently counted isolated sEVs. EVics enabled early observations of sEV secretion in various cell lines and reduced the cost of evaluating the inhibitory effect of sEV inhibitors by 20-fold. Using EVics, sEVs concentrations and sEV PD-L1 were monitored in a 23-day cancer mouse model, and 160 clinical samples were prepared and successfully applied to diagnosis. These results demonstrate that EVics could become an innovative system for novel findings in basic and applied studies in sEV research.

Super-resolution of digital rock images with hybrid attention multi-branch neural network

High-resolution digital rock images are vital in studying microstructure and fluid flow characterization of oil and gas reservoirs. Computed tomography (CT) is currently the primary tool to acquire digital rock images because of its non-destructiveness and three-dimensional imaging capability. However, the restriction of imaging capability of CT sets a compromise between image resolution and field of view (FOV). The Hybrid Attention Multi-branch Super-Resolution Network (HAMSR) proposed in this study aims to remedy this deficiency. HAMSR employs a limited discrete distribution to capture all possible high-order information and generate reliable high-resolution images. The network is built to prioritize the recovery of pores, cracks, and textures in digital rock images by introducing channel and spatial attention mechanisms. The structure together with modules of HAMSR have been designed and optimized to yield superior performance with fewer parameters and less training time. HAMSR and several contemporary advanced models are trained on the DeepRock-SR 2D dataset which contains lots of sandstone and carbonate images. It is evident that HAMSR-generated images have minimal noise, clearer edges, and more micropores. Quantitative calculations show that compared to traditional bicubic interpolation, HAMSR acquires a 2.669 dB and 1.376 dB increase in peak signal-to-noise ratio (PSNR) respectively on the test sets (27%–38% reduction in relative error). Furthermore, the high-resolution (HR) images reconstructed by HAMSR are generally consistent with real HR images concerning statistical characteristics such as porosity, pore size distribution, shape factor distribution, and two-point correlation function. This work provides a new super-resolution model to generate high-quality digital rock images for further analysis.

Research on the influence of measurement accuracy of the CCD moiré method

The CCD moiré method is an emerging tool for full-field measurements due to its simple optical setup and high resolution in topography and deformation measurements. It takes the periodic pixel on the image sensor of the digital camera as the reference grating, effectively “magnifies” the specimen grating pasted on the specimen surface. Different from ordinary moiré, CCD moiré has an image sensor that serves as both a reference grating and a recording pixel, its interval of parameters for measurement is narrow, and moiré quality is affected by multiple factors. This paper presents a systematic study of the factors affecting the accuracy of CCD moiré measurement. Firstly, a theoretical deduction of the formation process and measurement principle of CCD moiré is conducted, and then four classical single-moiré image-based phase extraction algorithms are summarized. Secondly, the imaging model of CCD moiré is established and all the factors that affect the accuracy of CCD moiré are studied, including lens aperture-coefficient FN, the pith of grating (ps, pr), pixel's duty-ratio dr, and camera acquisition mode. The results have identified the following optimal parameter interval: (1) determining the optimal pitch-ratio (n, or, ps / pr) is crucial for achieving high image quality and measurement accuracy, and the recommended interval isn∈[1.1,1.25]; (2) the pixel's duty-ratio is preferred to be set atdr∈[0.5,0.7], as it enables a favorable tradeoff between the image signal-to-noise ratio(SNR) and image contrast; (3) the recommended aperture-coefficient is FN∈[4,8]; (4) the Skip and Normal camera acquisition mode is preferred based on actual measurement situations. Finally, the measurement accuracy and advantages of the CCD moiré method are effectively verified through two experiments: one on four-point bending in comparison with the DIC method, and the other on the ship vertical arch displacement measurement under a large field of view (FOV).

CTIS spectral image reconstruction technology based on slit-scanning architecture

The computed tomography imaging spectrometer (CTIS) is a snapshot imaging spectrometer, excelling in dynamic detection tasks. It can capture two-dimensional spatial information and spectrally compressed information of a target within a single exposure time. However, traditional CTIS image reconstruction algorithms suffer from missing-cone problem, which reduces the accuracy of spectral reconstruction. In recent years, deep learning has been applied to CTIS spectral image reconstruction, significantly improving spectral reconstruction accuracy compared to traditional algorithms. However, due to the missing-cone problem, it is difficult to accurately recover the truth of spectral data cube in the real scene. Currently, most CTIS neural network reconstruction models are trained using simulated datasets of spectral data cubes and diffractive images. Because these data differ significantly from real data under actual application conditions, the established models may not be effectively applicable to real-world scenes. Therefore, we propose a new CTIS system based on slit-scanning architecture utilizing an adjustable slit aperture to obtain the real spectral data cube of the target while maintaining the simplicity of the CTIS structure. By limiting the field of view (FOV) through the slit, the area of diffraction overlap can be reduced, thereby enhancing the accuracy of CTIS spectral reconstruction using the expectation-maximization (EM) algorithm. This architecture allows us to obtain accurate spectral cubes that match the CTIS diffractive image of real-world scenes, providing a real dataset for training the reconstruction network. A prototype has been built to demonstrate the feasibility of our proposed solution. Furthermore, we also constructed a residual network based on multi-scale and attention mechanism. This network is trained using a combination of simulated and real spectral imaging data. Compared to the reconstruction performance of the EM algorithm and convolutional neural networks, our approach demonstrates superior spectral reconstruction accuracy, validating the importance of real spectral data in CTIS spectral reconstruction tasks.

Multiple-view 3D digital image correlation based on pseudo-overlapped imaging.

The restriction of the field of view (FOV) enlargement and spatial resolution increase during optical monitoring was investigated. Traditional optical instruments usually have a fixed FOV in one test; thus, they have low accuracy for small samples under large motions/deformations. To improve the spatial resolution in a relatively large FOV of an optical instrument, a multiple-view 3D digital image correlation (3D-DIC) method based on pseudo-overlapped imaging is proposed. Using a set of optical components, pseudo-overlapped imaging can image two FOVs with the same camera, which converts one pair of cameras to four virtual cameras. Each virtual camera has the same whole pixels. Therefore, compared with the conventional 3D-DIC system, the proposed method simultaneously enlarges FOVs and increases spatial resolutions by two times. The efficiency, accuracy, and feasibility of the technique were validated through experiments.

High-speed forward-viewing optical coherence tomography probe based on Lissajous sampling and sparse reconstruction.

We present a novel endoscopy probe using optical coherence tomography (OCT) that combines sparse Lissajous scanning and compressed sensing (CS) for faster data collection. This compact probe is only 4 mm in diameter and achieves a large field of view (FOV) of 2.25 mm2 and a 10 mm working distance. Unlike traditional OCT systems that use bulky raster scanning, our design features a dual-axis piezoelectric mechanism for efficient Lissajous pattern scanning. It employs compressive data reconstruction algorithms that minimize data collection requirements for efficient, high-speed imaging. This approach significantly enhances imaging speed by over 40%, substantially improving miniaturization and performance for endoscopic applications.

Dual-source dual-energy CT and deep learning for equivocal lymph nodes on CT images for thyroid cancer.

This study investigated the diagnostic performance of dual-energy computed tomography (CT) and deep learning for the preoperative classification of equivocal lymph nodes (LNs) on CT images in thyroid cancer patients. In this prospective study, from October 2020 to March 2021, 375 patients with thyroid disease underwent thin-section dual-energy thyroid CT at a small field of view (FOV) and thyroid surgery. The data of 183 patients with 281 LNs were analyzed. The targeted LNs were negative or equivocal on small FOV CT images. Six deep-learning models were used to classify the LNs on conventional CT images. The performance of all models was compared with pathology reports. Of the 281 LNs, 65.5% had a short diameter of less than 4 mm. Multiple quantitative dual-energy CT parameters significantly differed between benign and malignant LNs. Multivariable logistic regression analyses showed that the best combination of parameters had an area under the curve (AUC) of 0.857, with excellent consistency and discrimination, and its diagnostic accuracy and sensitivity were 74.4% and 84.2%, respectively (p < 0.001). The visual geometry group 16 (VGG16) based model achieved the best accuracy (86%) and sensitivity (88%) in differentiating between benign and malignant LNs, with an AUC of 0.89. The VGG16 model based on small FOV CT images showed better diagnostic accuracy and sensitivity than the spectral parameter model. Our study presents a noninvasive and convenient imaging biomarker to predict malignant LNs without suspicious CT features in thyroid cancer patients. Our study presents a deep-learning-based model to predict malignant lymph nodes in thyroid cancer without suspicious features on conventional CT images, which shows better diagnostic accuracy and sensitivity than the regression model based on spectral parameters. Many cervical lymph nodes (LNs) do not express suspicious features on conventional computed tomography (CT). Dual-energy CT parameters can distinguish between benign and malignant LNs. Visual geometry group 16 model shows superior diagnostic accuracy and sensitivity for malignant LNs.