Previous Image Research Articles

A Deep Cryptographic Framework for Securing the Healthcare Network from Penetration.

Ensuring the security of picture data on a network presents considerable difficulties because of the requirement for conventional embedding systems, which ultimately leads to subpar performance. It poses a risk of unauthorized data acquisition and misuse. Moreover, the previous image security-based techniques faced several challenges, including high execution times. As a result, a novel framework called Graph Convolutional-Based Twofish Security (GCbTS) was introduced to secure the images used in healthcare. The medical data are gathered from the Kaggle site and included in the proposed architecture. Preprocessing is performed on the data inserted to remove noise, and the hash 1 value is computed. Using the generated key, these separated images are put through the encryption process to encrypt what they contain. Additionally, to verify the user's identity, the encrypted data calculates the hash 2 values contrasted alongside the hash 1 value. Following completion of the verification procedure, the data are restored to their original condition and made accessible to authorized individuals by decrypting them with the collective key. Additionally, to determine the effectiveness, the calculated results of the suggested model are connected to the operational copy, which depends on picture privacy.

Brain and eye movement dynamics track the transition from learning to memory-guided action

Learning never stops. As we navigate life, we continuously acquire and update knowledge to optimize memory-guided action, with a gradual shift from the former to the latter as we master our environment. How are these learning dynamics expressed in the brain and in behavioral patterns? Here, we devised a spatiotemporal image learning task ("Memory Arena") in which participants learn a set of 50 items to criterion across repeated exposure blocks. Critically, brief task-free periods between successive image presentations allowed us to assess multivariate electroencephalogram (EEG) patterns representing the previous and/or upcoming image identity, as well as anticipatory eye movements toward the upcoming image location. As expected, participants eventually met the performance criterion, albeit with different learning rates. During task-free periods, we were able to readily decode representations of both previous and upcoming image identities. Importantly though, decoding strength followed opposing slopes for previous vs. upcoming images across time, with a gradual decline of evidence for the previous image and a gradual increase of evidence for the upcoming image. Moreover, the ratio of upcoming vs. previous image evidence directly followed behavioral learning rates. Finally, eye movement data revealed that participants increasingly used the task-free period to anticipate upcoming image locations, with target-precision slopes paralleling both behavioral performance measures as well as EEG decodability of the upcoming image across time. Together, these results unveil the neural and behavioral dynamics underlying the gradual transition from learning to memory-guided action.

X-Tracker: Automated Analysis of Xenopus Tadpole Visual Avoidance Behavior.

Xenopus laevis tadpoles exhibit an avoidance behavior when they encounter a moving visual stimulus. A visual avoidance event occurs when a moving object approaches the eye of a free-swimming animal at an approximately 90-degree angle and the animal turns in response to the encounter. Analysis of this behavior requires tracking both the free-swimming animal and the moving visual stimulus both prior to and after the encounter. Previous automated tracking software does not discriminate the moving animal from the moving stimulus, requiring time-consuming manual analysis. Here we present X-Tracker, an automated behavior tracking code that can detect and discriminate moving visual stimuli and free-swimming animals and score encounters and avoidance events. X-Tracker is as accurate as human analysis without the human time commitment. We also present software improvements to our previous visual stimulus presentation and image capture that optimize videos for automated analysis, and hardware improvements that increase the number of animal-stimulus encounters. X-Tracker is a high throughput, unbiased, and significant time-saving analysis system that will greatly facilitate visual avoidance behavior analysis of Xenopus laevis tadpoles, and potentially other free-swimming organisms. The tool is available at https://github.com/ClineLab/Tadpole-Behavior-Automation.

Reflections in a Golden Eye: Exploring the Photographic Art of Margaret Nolan and Shirley Eaton

This article offers a new perspective on the much-discussed phenomenon of the ‘Bond girl’ through an exploration of photographic artwork produced in later life by two notable Bond girls: the two ‘golden girls’ of Goldfinger (1964), Margaret Nolan, who appeared in the film’s credit sequence, and Shirley Eaton, who played the character who suffers skin suffocation by being painted gold. It is a remarkable coincidence that both women turned to photographic artwork as a way of reflecting back on their careers and on their youthful positioning as sex symbols, exemplified by their highly publicised appearances in Goldfinger. This article places their 1960s iconicity into dialogue with their retrospective, reflective artwork which is by turns critical, celebratory and ambivalent about their past lives as Bond girls. Nolan’s feminist-inflected photo-collages render her previous glamorous image uncanny through unsettling and subversive juxtapositions whereas Eaton’s digital remixing of old images alongside contemporary self-portraiture occupy the territory of star glamour more comfortably but extend it by showing the golden girl as older woman. Nolan and Eaton’s artwork provides valuable insights into the experiential dimensions of having been a Bond girl, with their creative response through art presenting certain complexities and nuances that written or spoken accounts might not access in quite the same way. In that respect, this article, alongside its specific insights into Bond girls, demonstrates the value for researchers in engaging with such material.

Automatic ploidy prediction and quality assessment of human blastocysts using time-lapse imaging

Assessing fertilized human embryos is crucial for in vitro fertilization, a task being revolutionized by artificial intelligence. Existing models used for embryo quality assessment and ploidy detection could be significantly improved by effectively utilizing time-lapse imaging to identify critical developmental time points for maximizing prediction accuracy. Addressing this, we develop and compare various embryo ploidy status prediction models across distinct embryo development stages. We present BELA, a state-of-the-art ploidy prediction model that surpasses previous image- and video-based models without necessitating input from embryologists. BELA uses multitask learning to predict quality scores that are thereafter used to predict ploidy status. By achieving an area under the receiver operating characteristic curve of 0.76 for discriminating between euploidy and aneuploidy embryos on the Weill Cornell dataset, BELA matches the performance of models trained on embryologists’ manual scores. While not a replacement for preimplantation genetic testing for aneuploidy, BELA exemplifies how such models can streamline the embryo evaluation process.

Solving East Nile Delta's imaging challenges through new ocean-bottom-node acquisition and advanced imaging

The East Nile Delta, located offshore Egypt, is a challenging area for seismic imaging due to the presence of substantial velocity heterogeneity created by the complex Messinian unconformity and overlaying fault structures. Existing towed-streamer (TS) seismic data have been pushed to their technical limit but remain insufficient for adequate representation of the subsurface. Previous studies have shown that a step change in imaging is possible if advanced acquisition, processing, and velocity model building techniques are employed. Atoll Field is a candidate to benefit from such improvements in seismic data. Results show that ocean-bottom-node acquisition provides improved spatial sampling and higher resolution. Acquiring seismic data with offsets of more than 30 km facilitated the generation of a robust high-frequency velocity model that better captures subsurface complexity. This paper outlines the acquisition and processing, with a focus on the velocity model building process. The generated subsurface model is effectively used for imaging with conventional migration algorithms and full-waveform-inversion-derived reflectivity. Results show a significant step change from previous TS images at the Oligocene target, with channel features clearly imaged. This provides an improved understanding of the depositional systems and reduced depth uncertainty.

Optical image authentication and encryption scheme with computational ghost imaging

Most of the previous image encryption schemes overlook certifying the source of the encrypted images, thus the receiver may be misled by messages coming from unauthorized users or wrong sources. To simultaneously accomplish image encryption and source authentication of the encrypted images, an optical image authentication and encryption scheme is proposed by integrating computational ghost imaging (CGI) with quick response (QR) code. Firstly, the plaintext image is converted into a QR code and encrypted by a scrambling-diffusion structure. Then, the authentication image is enciphered by a CGI system and a diffusion structure. The intermediate resulting images are marked and embedded into the random matrices to withstand the differential attack. Moreover, the source of the encrypted images can be identified by the reconstructed authentication image and the nonlinear correlation coefficient (NCC) result. The authentication is considered successful if a sharp peak appears in the center of the NCC result. Numerical simulations show that the proposed scheme can not only withstand different attacks, but also avoid being misled by misinformation during the decryption process.

A Steep Slab at the Yap Trench Resulted from Subducting Oceanic Plateaus

Abstract Interaction of oceanic plateaus with trenches plays a vital role in subduction activities and tectonic evolutions. The Yap trench is a rare case of an oceanic plateau subduction system. However, the knowledge of the impacts of plateau-trench interaction on subduction activity is still insufficient, due to a lack of seismological observations. Using ocean-bottom seismometer data near the Yap trench from April 2016 to May 2017, we conduct seismicity analyses in the Yap subduction zone by utilizing a machine-learning algorithm and matched-filter detections. The pattern of seismicity in the Yap trench exhibits characteristics similar to typical active subduction zones. The seismicity delineates a steep subducting slab, which may have resulted from the blocking of the buoyant Caroline Plateau. The majority of earthquakes are shallower than 80 km in the event-detectable area in the Yap trench, much shallower than the potential slab depth of 350 km from the previous seismic tomography images.

Attribute-Driven Filtering: A new attributes predicting approach for fine-grained image captioning

Fine-grained image captioning with attribute information has garnered significant attention in the realms of computer vision and natural language processing, demanding precise and contextually relevant descriptions of visual content. While previous attribute-driven image captioning models have shown improvements, challenges remain, such as the independence of attribute predictors and caption generators and the semantic gap between images and attributes. Another common issue is the inclusion of all attributes at every time step, despite most attributes being irrelevant to the word currently being generated. This can divert the model’s attention toward erroneous semantic details, resulting in a performance decline. To address these issues, we propose a novel Attribute-Driven Filtering (ADF) captioning network designed to provide rich and nuanced descriptions. This model incorporates a unique Attribute Predictor Module (APM) that dynamically predicts the most pertinent attributes in accordance with the textual context, utilizing different attributes at various time steps. The novelty of this approach lies in recognizing that not all attributes hold equal relevance at each time step, and the APM filters out irrelevant attributes to generate precise and contextually relevant captions. Furthermore, this model features a fusion mechanism that integrates visual information from a conventional attention module with attribute information predicted by the APM, aiming to reduce the visual semantic gap between images and attributes. Extensive experimentation demonstrates that the ADF model outperforms advanced models, achieving impressive CIDEr-D scores of 72.0 (Flickr30K) and 123.3 (MS-COCO) through reinforcement learning optimization. It consistently surpasses baseline models across diverse evaluation metrics, highlighting its effectiveness and robustness.

Normalisation of evacuation under slow emergencies: The imposed story of ‘Beautiful New Hong Kong’

AbstractHong Kong, in contrast to its previous image as a glamourous global city, has recently been associated with negative keywords such as oppression, fear, violence and even human rights emergency, following the 2019 Anti‐Extradition Law Movement and later the implementation of National Security Law (NSL) in 2020. Meanwhile, the Hong Kong Special Area Administration presented a post‐NSL policy direction as promising a ‘Beautiful New Hong Kong’. This paper is aimed at understanding how Hongkongers have continuously lived with/against the imposed (re)writing of Hong Kong. We examined how Hongkongers have been taking initiatives to raise awareness about the ‘slow emergencies’ in Hong Kong and to counter it in various forms of ‘evacuation’, whether they are on the move or staying put in place, in order to pursue future‐making. We carried out a multi‐sited study on the two kinds of Hongkongers between January 2020 and August 2023. We talked to those who already left Hong Kong for Taiwan, the UK, Canada, and so forth, and with those who were debating about relocation and at the same time preparing for departure if necessary. We strategically read ‘evacuation’ in two senses: First, evacuation responds to emergencies and therefore by adopting ‘evacuation’ is itself a disagreement with the Beautiful New Hong Kong policy as curated by the state. Second, evacuation responds to geography of future and politics of simultaneity. We conceptualised ‘normalisation of evacuation’ to understand the future‐making behind the move in Hong Kong. The particular kind of evacuation discussed took shape in two forms, relocation elsewhere and reorganisation in situ, both of which, in our analysis, are demonstrating Hongkongers' agency in pursuing geographies of future beyond the state‐led agenda.

El Gordo needs El Anzuelo: Probing the structure of cluster members with multi-band extended arcs in JWST data

Gravitational lensing by galaxy clusters involves hundreds of galaxies over a large redshift range and increases the likelihood of rare phenomena (supernovae, microlensing, dark substructures, etc.). Characterizing the mass and light distributions of foreground and background objects often requires a combination of high-resolution data and advanced modeling techniques. We present the detailed analysis of a prominent quintuply imaged dusty star-forming galaxy ($ mainly lensed by three members of the massive galaxy cluster ACT-CL\,J0102$-$4915, also known as d =0.87$). We leverage JWST/NIRCam images, which contain lensing features that were unseen in previous HST images, using a Bayesian, multi-wavelength, differentiable and GPU-accelerated modeling framework that combines (lens modeling) and (field model and inference) software packages. For one of the deflectors, we complement lensing constraints with stellar kinematics measured from VLT/MUSE data. In our lens model, we explicitly include the mass distribution of the cluster, locally corrected by a constant shear field. We find that the two main deflectors (L1 and L2) have logarithmic mass density slopes steeper than isothermal, with $ L1 and $ L2 We argue that such steep density profiles can arise due to tidally truncated mass distributions, which we probe thanks to the cluster lensing boost and the strong asymmetry of the lensing configuration. Moreover, our three-dimensional source model captures most of the surface brightness of the lensed galaxy, revealing a clump with a maximum diameter of $400$ parsecs at the source redshift, visible at wavelengths $ rest mu m. Finally, we caution on using point-like features within extended arcs to constrain galaxy-scale lens models before securing them with extended arc modeling.

Validation of Contrast-Enhanced Mammography as Breast Imaging Modality Compared to Standard Mammography and Digital Breast Tomosynthesis.

Contrast-enhanced mammography (CEM) is a relatively new imaging technique that allows morphologic, anatomic and functional imaging of the breast. The aim of our study was to validate contrast-enhanced mammography (CEM) compared to mammography (MMG) and digital breast tomosynthesis (DBT) in daily clinical practice. This retrospective study included 316 consecutive patients who underwent MMG, DBT and CEM at the Centre for Prevention and Diagnosis of Chronic Diseases of Primorsko-goranska County. Two breast radiologists independently analyzed the image data, without available anamnestic information and without the possibility of comparison with previous images, to determine the presence of suspicious lesions and their morphological features according to the established criteria of the Breast Imaging Reporting and Data System (BI-RADS) lexicon. The diagnostic value of MMG, DBT and CEM was assessed by ROC analysis. The interobserver agreement was excellent. CEM showed higher diagnostic accuracy in terms of sensitivity and specificity compared to MMG and DBT, the reporting time for CEM was significantly shorter, and CEM findings resulted in a significantly lower proportion of equivocal findings (BI-RADS 0), suggesting fewer additional procedures. In conclusion, CEM achieves high diagnostic accuracy while maintaining simplicity, reproducibility and applicability in complex clinical settings.

Integrated Multi-Scale Aircraft Detection and Recognition with Scattering Point Intensity Adaptiveness in Complex Background Clutter SAR Images

Detecting aircraft targets in Synthetic Aperture Radar (SAR) images is critical for military and civilian applications. However, due to SAR’s special imaging mechanism, aircraft targets often consist of scattering points with large fluctuations in intensity. This often leads to the detector failing to detect weak scattering points. Not only that, previous SAR image aircraft-object-detection models have focused more on detecting and locating targets, with little emphasis on target recognition. This paper proposes a scattering-point-intensity-adaptive detection and recognition network (SADRN). In order to correctly detect the target area, we propose a Self-adaptive Bell-shaped Kernel (SBK) within the detector, which constructs a bell-shaped two-dimensional distribution centered on the target center, making the detection “threshold” for the target decrease from the center towards the periphery, reducing the missed alarms of weak scattering points at the edges of the target. To help the model adapt to multi-scale targets, we propose the FADLA-34 backbone network, aggregating information from feature maps across different scales. We also embed CBAM into the detector, which enhances the attention to the target area in the spatial dimension and strengthens the extraction of useful features in the channel dimension, reducing interference from the complex background clutter on object detection. Furthermore, to integrate detection and recognition, we introduce the multi-task head, which utilizes the three feature maps from the backbone network to generate the detection boxes and categories of the targets. Finally, the SADRN achieves superior detection and recognition performance on the SAR-AIRcraft-1.0, exceeding other mainstream methods. Visualization and analysis further confirm the effectiveness and superiority of the SADRN.

Tackling bisexual erasure: An explorative comparison of bisexual, gay and straight cisgender men’s body image

Previous body image research often grouped both gay and bisexual men into a single category: sexual minoritised men, limiting our understanding of how sexual identity influences body image. However, there is strong reason to believe that bisexual and gay men experience distinct body image concerns. Here, we explored motivations to alter one’s leanness and muscularity, as well as (dis)satisfaction with body fat, muscularity, height and penis size, and functionality appreciation across gay, bisexual, and straight cisgender men. We sampled 378 white participants aged 18 to 85 (nbisexual = 125, ngay = 128, nstraight = 125). We found that bisexual men were significantly less motivated to be lean and showed lower muscularity dissatisfaction relative to gay men but showed comparable levels to straight men. Our findings demonstrate that despite research perceiving the body image of bisexual and gay men as homogenous, they experience differences in their body image concerning leanness and muscularity dissatisfaction. Future body image research should incorporate this understanding by not artificially grouping bisexual and gay cisgender men and instead acknowledging the potential uniqueness in their experiences.

Radiographic assessment of the peri-implant site.

While peri-implant mucositis relies solely on clinical parameters, radiological assessment becomes indispensable for diagnosing peri-implantitis. Intraoral radiography, with its simplicity of application, low radiation exposure, and adequate representation of peri-implant structures, stands out as the standard of care for both immediate and follow-up assessments. Standardization by custom-made radiologic splints allows for excellent comparability with previous images and allows for the determination of even small changes in contour and density of the peri-implant bone. Furthermore, other radiographic modalities like panoramic radiography and cone beam computed tomography (CBCT) may provide useful features for specific patients and clinical cases while also showing innate limitations. Beyond the assessment of the marginal peri-implant bone level as the crucial parameter of clinical relevance, radiologic assessment may reveal various other findings related to the prosthetic restoration itself, the precision of its fit to the implant, and the peri-implant soft and hard tissues. Since such findings can be crucial for the assessment of peri-implant health and the implants' prognosis, a systematic diagnostic evaluation pathway for a thorough assessment is recommended to extract all relevant information from radiologic imaging. This article also provides an overview of the clinical and chronological indications for different imaging modalities in peri-implant issues.

Reproducibility Analysis of Radiomic Features on T2-weighted MR Images after Processing and Segmentation Alterations in Neuroblastoma Tumors.

Purpose To evaluate the reproducibility of radiomics features extracted from T2-weighted MR images in patients with neuroblastoma. Materials and Methods A retrospective study included 419 patients (mean age, 29 months ± 34 [SD]; 220 male, 199 female) with neuroblastic tumors diagnosed between 2002 and 2023, within the scope of the PRedictive In-silico Multiscale Analytics to support cancer personalized diaGnosis and prognosis, Empowered by imaging biomarkers (ie, PRIMAGE) project, involving 746 T2/T2*-weighted MRI sequences at diagnosis and/or after initial chemotherapy. Images underwent processing steps (denoising, inhomogeneity bias field correction, normalization, and resampling). Tumors were automatically segmented, and 107 shape, first-order, and second-order radiomics features were extracted, considered as the reference standard. Subsequently, the previous image processing settings were modified, and volumetric masks were applied. New radiomics features were extracted and compared with the reference standard. Reproducibility was assessed using the concordance correlation coefficient (CCC); intrasubject repeatability was measured using the coefficient of variation (CoV). Results When normalization was omitted, only 5% of the radiomics features demonstrated high reproducibility. Statistical analysis revealed significant changes in the normalization and resampling processes (P < .001). Inhomogeneities removal had the least impact on radiomics (83% of parameters remained stable). Shape features remained stable after mask modifications, with a CCC greater than 0.90. Mask modifications were the most favorable changes for achieving high CCC values, with a radiomics features stability of 70%. Only 7% of second-order radiomics features showed an excellent CoV of less than 0.10. Conclusion Modifications in the T2-weighted MRI preparation process in patients with neuroblastoma resulted in changes in radiomics features, with normalization identified as the most influential factor for reproducibility. Inhomogeneities removal had the least impact on radiomics features. Keywords: Pediatrics, MR Imaging, Oncology, Radiomics, Reproducibility, Repeatability, Neuroblastic Tumors Supplemental material is available for this article. © RSNA, 2024 See also the commentary by Safdar and Galaria in this issue.

JWST and ALMA Discern the Assembly of Structural and Obscured Components in a High-redshift Starburst Galaxy

We present observations and analysis of the starburst PACS-819 at z = 1.45 (M * = 1010.7 M ⊙), using high-resolution (0.″1; 0.8 kpc) Atacama Large Millimeter/submillimeter Array (ALMA) and multiwavelength JWST images from the COSMOS-Web program. Dissimilar to Hubble Space Telescope (HST) ACS images in the rest-frame UV, the redder NIRCam and MIRI images reveal a smooth central mass concentration and spiral-like features, atypical for such an intense starburst. Through dynamical modeling of the CO (J = 5–4) emission with ALMA, PACS-819 is rotation dominated and thus consistent with having a disk-like nature. However, kinematic anomalies in CO and asymmetric features in the bluer JWST bands (e.g., F150W) support a more disturbed nature likely due to interactions. The JWST imaging further enables us to map the distribution of stellar mass and dust attenuation, thus clarifying the relationships between different structural components not discernible in the previous HST images. The CO (J = 5–4) and far-infrared dust continuum emission are cospatial with a heavily obscured starbursting core (<1 kpc) that is partially surrounded by much less obscured star-forming structures including a prominent arc, possibly a tidally distorted dwarf galaxy, and a massive clump (detected in CO), likely a recently accreted low-mass satellite. With spatially resolved maps, we find a high molecular gas fraction in the central area reaching ∼3 (M gas/M *) and short depletion times (M gas/SFR ∼ 120 Myr, where SFR is star formation rate) across the entire system. These observations provide insights into the complex nature of starbursts in the distant Universe and underscore the wealth of complementary information from high-resolution observations with both ALMA and JWST.

High‐density offshore seismic exploration with an optical fibre towed streamer based on distributed acoustic sensing: Concept and application

AbstractSeismic technique is widely used to image the subsurface geology for oil and gas exploration. The image quality depends on the spatial sampling density. However, it is challenging and expensive to acquire high‐density seismic data, particularly in the marine environment. Distributed acoustic sensing data are increasingly used in data acquisition because of their low cost and dense spatial sampling. Here, we present a novel type of high‐density towed streamer based on distributed acoustic sensing technology and report the results of a sea trial. This sea trial was conducted in a gas hydrate province as the major driver to develop this technique is to better characterize gas hydrate deposits. Throughout the experiment, several high‐quality datasets were obtained, and parameters like source energies and filler materials were examined. The trace interval of distributed acoustic sensing streamer data reaches 1 m, which is a significant improvement over the usual 3.125 or 6.25 m in the conventional towed streamer. A detailed analysis was carried out from three different perspectives: amplitude, noise and frequency. One of the datasets was further processed following a routine workflow to obtain the final image. Though direct comparison with the image obtained by a conventional towed streamer along a coincident line is not available, the comparison with the previous image from a nearby line shows the improvement in resolution. The final image is of good quality and the presence of gas hydrate could be inferred. The sea trial results demonstrate the feasibility of the use of a distributed acoustic sensing optical fibre streamer in acquiring high‐density seismic data in the marine environment.

Matte anything: Interactive natural image matting with segment anything model

Natural image matting algorithms aim to predict the transparency map (alpha-matte) with the trimap guidance. However, the production of trimap often requires significant labor, which limits the widespread application of matting algorithms on a large scale. To address the issue, we propose Matte Anything model (MatAny), an interactive natural image matting model that could produce high-quality alpha-matte with various simple hints. The key insight of MatAny is to generate pseudo trimap automatically with contour and transparency prediction. In our work, we leverage vision foundation models to enhance the performance of natural image matting. Specifically, we use the segment anything model to predict high-quality contour with user interaction and an open-vocabulary detector to predict the transparency of any object. Subsequently, a pre-trained image matting model generates alpha mattes with pseudo trimaps. MatAny is the interactive matting algorithm with the most supported interaction methods and the best performance to date. It consists of orthogonal vision models without any additional training. We evaluate the performance of MatAny against several previous image matting algorithms. MatAny has 58.3% improvement on MSE and 40.6% improvement on SAD compared to the previous image matting methods with simple guidance, achieving new state-of-the-art (SOTA) performance. The source codes and pre-trained models are available at https://github.com/hustvl/Matte-Anything.

Application of MR images in radiotherapy planning for brain tumor based on deep learning

Purpose Explore the function and dose calculation accuracy of MRI images in radiotherapy planning through deep learning methods. Methods 131 brain tumor patients undergoing radiotherapy with previous MR and CT images were recruited for this study. A new series of MRI from the aligned MR was firstly registered to CT images strictly using MIM software and then resampled. A deep learning method (U-NET) was used to establish a MRI-to-CT conversion model, for which 105 patient images were used as the training set and 26 patient images were used as the tuning set. Data from additional 8 patients were collected as the test set, and the accuracy of the model was evaluated from a dosimetric standpoint. Results Comparing the synthetic CT images with the original CT images, the difference in dosimetric parameters D98, D95, D2 and Dmean of PTV in 8 patients was less than 0.5%. The gamma passed rates of PTV and whole body volume were: 1%/1 mm: 93.96%±6.75%, 2%/2 mm: 99.87%±0.30%, 3%/3 mm: 100.00%±0.00%; and 1%/1 mm: 99.14%±0.80%, 2%/2 mm: 99.92%±0.08%, 3%/3 mm: 99.99%±0.01%. Conclusion MR images can be used both in delineation and treatment efficacy evaluation and in dose calculation. Using the deep learning way to convert MR image to CT image is a viable method and can be further used in dose calculation.