Landmarks In Images Research Articles

A DCNN system based on an iterative method for automatic landmark detection in cephalometric X-ray images

Cephalometric analysis is a commonly used method in modern orthodontic clinical diagnosis and treatment. The traditional way of manually marking cephalometric landmarks takes a long time, and it is difficult to obtain stable detection accuracy because of the uneven professionalism of doctors. In order to efficiently and accurately locate the landmarks of X-ray cephalograms, we propose a deep convolutional neural network (DCNN) system for automatic landmark detection in cephalometric X-ray images. And the system incorporates an iterative method while introducing the ideas of transfer learning and data enhancement. We first preprocess and expand the dataset of the cephalometric X-ray images. Then we use the preprocessed image to train a deep convolutional neural network and obtain preliminary prediction areas for various landmarks. Next, for each type of landmark, a detection process based on the iterative method is used to gradually reduce the prediction range in five iterations, which achieves more precise landmark positioning. Finally, the system restores the final predicted position in the original image by reversing the predicted coordinates. Our training set and test set proposed in the 2015 IEEE International Symposium on Biomedical Imaging each contain 150 cephalometric X-ray images. The accuracy of the system predicting 19 pairs of landmarks in the image within the error range of two millimeters reaches 87.51% on average. The experimental results and comparisons show that the iterative method does improve the accuracy and stability of location prediction and our cephalometric image landmark detector is effective in locating landmarks.

Artificial intelligence clustering of adult spinal deformity sagittal plane morphology predicts surgical characteristics, alignment, and outcomes.

AI algorithms have shown promise in medical image analysis. Previous studies of ASD clusters have analyzed alignment metrics-this study sought to complement these efforts by analyzing images of sagittal anatomical spinopelvic landmarks. We hypothesized that an AI algorithm would cluster preoperative lateral radiographs into groups with distinct morphology. This was a retrospective review of a multicenter, prospectively collected database of adult spinal deformity. A total of 915 patients with adult spinal deformity and preoperative lateral radiographs were included. A 2 × 3, self-organizing map-a form of artificial neural network frequently employed in unsupervised classification tasks-was developed. The mean spine shape was plotted for each of the six clusters. Alignment, surgical characteristics, and outcomes were compared. Qualitatively, clusters C and D exhibited only mild sagittal plane deformity. Clusters B, E, and F, however, exhibited marked positive sagittal balance and loss of lumbar lordosis. Cluster A had mixed characteristics, likely representing compensated deformity. Patients in clusters B, E, and F disproportionately underwent 3-CO. PJK and PJF were particularly prevalent among clusters A and E. Among clusters B and F, patients who experienced PJK had significantly greater positive sagittal balance than those who did not. This study clustered preoperative lateral radiographs of ASD patients into groups with highly distinct overall spinal morphology and association with sagittal alignment parameters, baseline HRQOL, and surgical characteristics. The relationship between SVA and PJK differed by cluster. This study represents significant progress toward incorporation of computer vision into clinically relevant classification systems in adult spinal deformity. Diagnostic: individual cross-sectional studies with the consistently applied reference standard and blinding.

Case study of Service Design for KNUT Dulle-gil

Recently, a landmark representing the local community is formed by giving meaning to a building or place based in the area. By giving artificial meaning along with natural conditions, it plays a role in attracting tourists as well as residents. Therefore, this study intends to study service design targeting the Dulle-gil of Korea Transportation University located in Chungju. For this, a new design method was studied by identifying the current condition of the Dulle-gil and analyzing the inconvenience factors. Dullegil is a space where facilities with the functions of exercise, rest, and walking harmoniously harmonize with the natural environment. It is a sufficient space to seek satisfaction, gain experiences, and recharge. In particular, it can become a medium that can instill a positive image of Korea and schools to foreign students, and furthermore, it will become an intangible asset that can serve as an ambassador for future Korea. Therefore, by researching and suggesting the design elements that make up the Dullegil of Korea Transportation University, the purpose is to build a landmark image of Korea Transportation University as a local attraction that not only students, faculty, but also residents can enjoy.

Juvenile Chinook Salmon Weight Prediction Using Image‐Based Morphometrics

AbstractWe developed an empirical weight prediction model for juvenile Chinook Salmon Oncorhynchus tshawytscha ranging from 33–113 mm FL and 0.35–14.86 g using morphometrics analysis on images that were collected from live fish in the field. This method relies on consumer‐grade hardware and free software and addresses the issue of a lack of valuable weight and condition data for life history and survival modeling. A blind test revealed that the method was relatively precise, with a mean absolute error of 0.075 g and a mean absolute percentage of error of 2.86%, and unbiased, with a mean error of −0.001 g. Given the inherent variability in field wet‐weight measurements that are estimated, with an average standard deviation of 0.030 g and CV of 1.02% from a blind repeated‐measurements experiment, the observed level of predictive error suggests that the method is an adequate substitute for measuring field weight when conditions (i.e., boat sampling) preclude conventional scale taring and accurate weight measurement. Additionally, this method is useful in an automated data‐quality‐control workflow to identify erroneous length and/or weight measurements. The image landmark placement closely follows previous work on juvenile Chinook Salmon and can be incorporated with existing methodologies such as stock analysis and phenotypic differentiation to increase the available biometric information.

A novel machine learning-based algorithm to identify and classify lesions and anatomical landmarks in colonoscopy images.

Computer-aided diagnosis (CAD)-based artificial intelligence (AI) has been shown to be highly accurate for detecting and characterizing colon polyps. However, the application of AI to identify normal colon landmarks and differentiate multiple colon diseases has not yet been established. We aimed to develop a convolutional neural network (CNN)-based algorithm (GUTAID) to recognize different colon lesions and anatomical landmarks. Colonoscopic images were obtained to train and validate the AI classifiers. An independent dataset was collected for verification. The architecture of GUTAID contains two major sub-models: the Normal, Polyp, Diverticulum, Cecum and CAncer (NPDCCA) and Narrow-Band Imaging for Adenomatous/Hyperplastic polyps (NBI-AH) models. The development of GUTAID was based on the 16-layer Visual Geometry Group (VGG16) architecture and implemented on Google Cloud Platform. In total, 7838 colonoscopy images were used for developing and validating the AI model. An additional 1273 images were independently applied to verify the GUTAID. The accuracy for GUTAID in detecting various colon lesions/landmarks is 93.3% for polyps, 93.9% for diverticula, 91.7% for cecum, 97.5% for cancer, and 83.5% for adenomatous/hyperplastic polyps. A CNN-based algorithm (GUTAID) to identify colonic abnormalities and landmarks was successfully established with high accuracy. This GUTAID system can further characterize polyps for optical diagnosis. We demonstrated that AI classification methodology is feasible to identify multiple and different colon diseases.

Indoor surveillance video based feature recognition for pedestrian dead reckoning

Driven by the connection of everything, various electronic devices and sensors have been applied to the Internet of Things (IoT) as sources of information. However, indoor surveillance cameras, which are a quality resource, are only used in the field of security monitoring. Based on the development of indoor surveillance camera about image features, this study proposes a positioning method of pedestrian dead reckoning (PDR) based on image feature recognition of indoor surveillance video. In this study, the images collected by traditional indoor surveillance cameras are used as the data source. By extracting the image features to establish an image landmark database, geographic coordinates are allocated to each feature image of the landmark database in order to establish indoor location benchmark points. These image landmarks are combined with pedestrian dead reckoning to obtain the accurate position estimation. The experimental results show that the proposed indoor surveillance video based feature recognition for pedestrian dead reckoning can effectively use the indoor camera resources as a position benchmark in order to achieve higher pedestrian positioning accuracy. Meanwhile, this strategy of establishing a landmark database based on existing monitoring videos effectively utilizes existing indoor monitoring resources and promotes the application of traditional surveillance cameras.

The use of an O-arm in endonasal endoscopic operations of the skull base

BackgroundEndoscopic endonasal transsphenoidal approaches are broadly used nowadays for a vast spectrum of pathologies sited in the anterior and middle cranial fossa. The usage of neuronavigation systems (neuronavigation) in these surgeries is crucial for improving orientations deeply inside the skull and increasing patient safety.MethodsThe aim of this study was to assess the use of optical neuronavigation, together with an intraoperative O-arm O2 imaging system, in a group of patients with hypophyseal adenoma that underwent a transnasal transsphenoidal surgery, and correlate the accuracy and its deviation during the navigational process against the use of conventional neuronavigation that uses preoperative MRI and CT scans. The overall group consisted of six patients, between 39 and 78 years old, with a diagnosis of hypophyseal adenoma. Patients were treated with an endoscopic transsphenoidal technique and all of them underwent preoperative MRI and CT scans of the brain. These images were used in the neuronavigation system StealthStation S7® during the surgery, where we defined two bony anatomical landmarks, such as a vomer or the origin of an intrasphenoidal septum, in each operated patient. The tip of the navigational instrument, under endoscopic control, pointed to these landmarks and the distance between the tip and the bony structure was measured on the neuronavigation system. Afterwards, intraoperative 3D x-ray imaging was performed via the mobile system O-arm O2® system with automatic transfer into the navigational system. Under endoscopic guidance, we localized the identical bony anatomical landmarks used in the previous measurement and re-measured the distance between the tip and bony landmark in images acquired by the O-arm. The results of both measurements were statistically compared.ResultsThe mean error of accuracy during conventional neuronavigation with usage of preoperative CT and MRI scans was 2.65 mm. During the neuronavigation, with utilization of intraoperative 3D O-arm images, the mean error of accuracy 0 mm. These mean errors of accuracy (both measurement methods were compared by nonparametric Wilcoxon test) had a statistically significant difference (p = 0.043).ConclusionsBased on this preliminary clinical study, we conclude that the O-arm is capable of providing intraoperative x-ray 3D images in sufficient spatial resolution in a clinically feasible acquisition. The mean error of accuracy during intraoperative navigation, based on 3D O-arm scans at the skull base, is significantly lower compared to the usage of navigation using conventional presurgical CT and MRI images. This suggests the suitability of this method for utilization during endoscopic endonasal skull base approaches.

A survey on Pose Estimation using Deep Convolutional Neural Networks

Human pose estimation is a technique, which identifies the human body’s landmarks in images and videos. Human pose estimation can be divided into single person pose estimation and multi-person pose estimation, also an estimated human poses in crowded places as well as in videos. Depends upon the application such as activity recognition, Animation, Sports, Augmented reality, etc., Pose estimation output can be in 2D or 3D coordinate format. 3D pose estimation is estimated considering joint angles in 2D. Some challenges like small and barely visible joints, strong articulations, occlusions, clothing, and lighting changes increase difficulty in estimating pose. Remarkable progress has gained in the field of human pose estimation using Deep learning-based CNN models. In this paper, we compare and summarized various deep learning models for pose estimation of a single person and multi-person.

Diversifying Inference Path Selection: Moving-Mobile-Network for Landmark Recognition.

Deep convolutional neural networks have largely benefited computer vision tasks. However, the high computational complexity limits their real-world applications. To this end, many methods have been proposed for efficient network learning, and applications in portable mobile devices. In this paper, we propose a novel Moving-Mobile-Network, named M2Net, for landmark recognition, equipped each landmark image with located geographic information. We intuitively find that M2Net can essentially promote the diversity of the inference path (selected blocks subset) selection, so as to enhance the recognition accuracy. The above intuition is achieved by our proposed reward function with the input of geo-location and landmarks. We also find that the performance of other portable networks can be improved via our architecture. We construct two landmark image datasets, with each landmark associated with geographic information, over which we conduct extensive experiments to demonstrate that M2Net achieves improved recognition accuracy with comparable complexity.

Multiple-Clothing Detection and Fashion Landmark Estimation Using a Single-Stage Detector

Fashion image analysis has attracted significant research attention owing to the availability of large-scale fashion datasets with rich annotations. However, existing deep learning models for fashion datasets often have high computational requirements. In this study, we propose a new model suitable for low-power devices. The proposed network is a one-stage detector that rapidly detects multiple cloths and landmarks in fashion images. The network is designed as a modification of the EfficientDet originally proposed by Google Brain. The proposed network simultaneously trains the core input features with different resolutions and applies compound scaling to the backbone feature network. The bounding box/class/landmark prediction networks maintain the balance between the speed and accuracy. Moreover, a low number of parameters and low computational cost make it efficient. Without image preprocessing, we achieved 0.686 mean average precision (mAP) in the bounding box detection and 0.450 mAP in the landmark estimation on the DeepFashion2 validation dataset with an inference time of 42 ms. We obtained optimal results in extensive experiments with loss functions and optimizers. Furthermore, the proposed method has the advantage of operating in low-power devices.

Language-based translation and prediction of surgical navigation steps for endoscopic wayfinding assistance in minimally invasive surgery

PurposeIn the context of aviation and automotive navigation technology, assistance functions are associated with predictive planning and wayfinding tasks. In endoscopic minimally invasive surgery, however, assistance so far relies primarily on image-based localization and classification. We show that navigation workflows can be described and used for the prediction of navigation steps.MethodsA natural description vocabulary for observable anatomical landmarks in endoscopic images was defined to create 3850 navigation workflow sentences from 22 annotated functional endoscopic sinus surgery (FESS) recordings. Resulting FESS navigation workflows showed an imbalanced data distribution with over-represented landmarks in the ethmoidal sinus. A transformer model was trained to predict navigation sentences in sequence-to-sequence tasks. The training was performed with the Adam optimizer and label smoothing in a leave-one-out cross-validation study. The sentences were generated using an adapted beam search algorithm with exponential decay beam rescoring. The transformer model was compared to a standard encoder-decoder-model, as well as HMM and LSTM baseline models.ResultsThe transformer model reached the highest prediction accuracy for navigation steps at 0.53, followed by 0.35 of the LSTM and 0.32 for the standard encoder-decoder-network. With an accuracy of sentence generation of 0.83, the prediction of navigation steps at sentence-level benefits from the additional semantic information. While standard class representation predictions suffer from an imbalanced data distribution, the attention mechanism also considered underrepresented classes reasonably well.ConclusionWe implemented a natural language-based prediction method for sentence-level navigation steps in endoscopic surgery. The sentence-level prediction method showed a potential that word relations to navigation tasks can be learned and used for predicting future steps. Further studies are needed to investigate the functionality of path prediction. The prediction approach is a first step in the field of visuo-linguistic navigation assistance for endoscopic minimally invasive surgery.

Magnetic resonance imaging landmarks for preoperative localization of inferior medial genicular artery: a proof of concept analysis

Magnetic resonance imaging landmarks for preoperative localization of inferior medial genicular artery: a proof of concept analysis

Facial metrics generated from manually and automatically placed image landmarks are highly correlated

Research on social judgments of faces often investigates relationships between measures of face shape taken from images (facial metrics), and either perceptual ratings of the faces on various traits (e.g., attractiveness) or characteristics of the photographed individual (e.g., their health). A barrier to carrying out this research using large numbers of face images is the time it takes to manually position the landmarks from which these facial metrics are derived. Although research in face recognition has led to the development of algorithms that can automatically position landmarks on face images, the utility of such methods for deriving facial metrics commonly used in research on social judgments of faces has not yet been established. Thus, across two studies, we investigated the correlations between four facial metrics commonly used in social perception research (sexual dimorphism, distinctiveness, bilateral asymmetry, and facial width to height ratio) when measured from manually and automatically placed landmarks. In the first study, in two independent sets of open access face images, we found that facial metrics derived from manually and automatically placed landmarks were typically highly correlated, in both raw and Procrustes-fitted representations. In study two, we investigated the potential for automatic landmark placement to differ between White and East Asian faces. We found that two metrics, facial width to height ratio and sexual dimorphism, were better approximated by automatic landmarks in East Asian faces. However, this difference was small, and easily corrected with outlier detection. These data validate the use of automatically placed landmarks for calculating facial metrics to use in research on social judgments of faces, but we urge caution in their use. We also provide a tutorial for the automatic placement of landmarks on face images.

P37. Artificial intelligence clustering of adult spinal deformity morphology predicts surgical characteristics, alignment, and outcomes

BACKGROUND CONTEXT AI algorithms have shown substantial promise in medical image analysis. Previous studies of ASD clusters have analyzed alignment metrics – this study sought to complement these efforts by analyzing images of anatomical landmarks. PURPOSE We hypothesized that a neural-network-based artificial intelligence (AI) algorithm would cluster preoperative lateral radiographs of into groups with distinct morphology. STUDY DESIGN/SETTING Retrospective cohort study. PATIENT SAMPLE A total of 915 patients with adult spinal deformity and preoperative lateral radiographs. OUTCOME MEASURES Schwab modifiers for SVA and PI-LL, three-column osteotomy, upper instrumented vertebrae, baseline Oswestry Disability Index, and 2-year likelihood of reaching MCID in ODI (set at -12.8). Proximal junctional kyphosis and proximal junctional failure were defined using previously published radiographic criteria. METHODS Vertebral locations for C3-L5, sacral endplate, and femoral heads were measured on lateral radiographs. Pixel locations were used to create a black-and-white overlay to the image, which was subsequently standardized for size and position using the femoral heads and sacral endplate. These images were used to train a self-organizing map (SOM). SOMs are a form of artificial neural network frequently employed in unsupervised classification tasks. RESULTS In total, 915 preoperative lateral radiographs were analyzed. A 2 × 3, toroidal SOM was trained. The mean spine shape was plotted for each cluster. Alignment, surgical characteristics, and outcomes were compared between clusters. Clusters C and D exhibited a particularly high proportion of patients with optimal (ie, modifier 0) values of PI-LL (65.0% and 68.5%) and SVA (72.8% and 53.1%). Conversely, clusters B, E, and F tended to have poor (ie, modifier ++) PI-LL (74.8%, 66.9%, and 74.6%) and SVA (75.5%, 48.6%, and 58.7%). 3-CO was most common among cluster A (26.8%), cluster B (32.6%), and cluster F (32.7%). UIV at T7-T12 was most common among cluster B (51.1%) and cluster F (60.3%). ODI CONCLUSIONS This study developed a self-organizing map that clustered preoperative lateral radiographs of ASD patients into groups with highly distinct overall spinal morphology. These clusters predicted alignment, surgical characteristics, and HRQOL. Further studies of this classification approach will expand to compare pre- and postoperative radiographs. FDA DEVICE/DRUG STATUS This abstract does not discuss or include any applicable devices or drugs.

Automated Cephalometric Landmark Localization using a Coupled Shape Model

Abstract Cephalometric analysis is an important method in orthodontics for the diagnosis and treatment of patients. It is performed manually in clinical practice, therefore automation of this time consuming task would be of great assistance. In order to provide dentists with such tools, a robust and accurate identification of the necessary landmarks is required. However, poor image quality of lateral cephalograms like low contrast or noise make this task difficult. In this paper, an approach for automatic landmark localization is presented and used to find 19 landmarks in lateral cephalometric images. An initial predicting of the individual landmark locations is done by using a 2-D coupled shape model to utilize the spatial relation between landmarks and other anatomical structures. These predictions are refined with a Hough Forest to determine the final landmark location. The approach achieves competitive performance with a successful detection rate of 70.24% on 250 images for the clinically relevant 2mm accuracy range.

Discovering informative features in large-scale landmark image collection

One of the key problems in image retrieval systems is the presence of irrelevant and noisy image content. Such content can cause significant confusion for the system. Therefore, there is a need to represent images with only informative features in order to improve the retrieval performance of the system or any subsequent process. In this article, we propose a method to identify the informative features in a large-scale image collection. We apply the frequent itemset mining (FIM) approach to extract visual features patterns from a list of images of the same object. Then, we generate feature pairs to measure the significance of each feature depending on the co-occurrence with its neighbouring features. In addition, we apply this feature selection technique to localise the landmark in the image. The performance of the proposed method is evaluated in terms of average precision (AP) on two benchmark data sets and found that it gives a comparable retrieval performance over the bag of visual words baseline system and the previous methods.

Landmark-free head pose estimation using fusion inception deep neural network

Head pose estimation is a typical computer vision task that has been applied to many helpful applications. It is still challenging due to occlusions, low resolution, and extreme pose changes. A method is proposed for landmark-free head pose estimation in a fusion inception deep neural network. The fusion inception network is developed to predict head pose angles without landmarks in RGB images only, which avoids the error caused by landmark location faults. A three-channel prediction module in the developed network is designed to perform classification and regression jointly. A data augmentation method is proposed for sample expansion to alleviate overfitting. A balance loss strategy is proposed to take place of cross-entropy for pose angle classification. The developed balance loss can be applied to deal with class imbalance. The proposed method has excellent performance in head pose estimation by comparison with state-of-the-art methods on some challenging datasets.

Effect of Device Configuration and Patient's Body Composition on the RF Heating and Nonsusceptibility Artifact of Deep Brain Stimulation Implants During MRI at 1.5T and 3T.

Patients with deep brain stimulation (DBS) implants have limited access to MRI due to safety concerns associated with RF-induced heating. Currently, MRI in these patients is allowed in 1.5T horizontal bore scanners utilizing pulse sequences with reduced power. However, the use of 3T MRI in such patients is increasingly reported based on limited safety assessments. Here we present the results of comprehensive RF heating measurements for two commercially available DBS systems during MRI at 1.5T and 3T. To assess the effect of imaging landmark, DBS lead configuration, and patient's body composition on RF heating of DBS leads during MRI at 1.5T and 3T. Phantom and ex vivo study. Gel phantoms and cadaver brain. 1.5T and 3T, T1 -weighted turbo spin echo. RF heating was measured at the tips of DBS leads implanted in brain-mimicking gel. Image artifact was assessed in a cadaver brain implanted with an isolated DBS lead. Descriptive. We observed substantial fluctuation in RF heating, mainly affected by phantom composition and DBS lead configuration, ranging from 0.14°C to 23.73°C at 1.5T, and from 0.10°C to 7.39°C at 3T. The presence of subcutaneous fat substantially altered RF heating at the electrode tips (3.06°C < ∆T < 19.05° C). Introducing concentric loops in the extracranial portion of the lead at the surgical burr hole reduced RF heating by up to 89% at 1.5T and up to 98% at 3T compared to worst-case heating scenarios. Device configuration and patient's body composition substantially altered the RF heating of DBS leads during MRI. Interestingly, certain lead trajectories consistently reduced RF heating and image artifact. Level of Evidence 1 Technical Efficacy Stage 1 J. MAGN. RESON. IMAGING 2021;53:599-610.

CONCEPT ON LANDMARK DETECTION IN ROAD SCENE IMAGES TAKEN FROM A TOP-VIEW CAMERA SYSTEM

Abstract. In this paper, we demonstrate the inclusion of a top-view camera system mounted on a city bus in an existing sensor setup. A novel sensor setup with five down-facing cameras is mounted on the roof of a MAN Lion’s City 12 city bus to extract landmarks in road scene images. Its positioning is validated by an exemplary detection of lane markings. The concept for further landmark detection with the help of the presented camera system is explained in this paper and sensor data fusion methods are proposed. Based on our previous findings (Albrecht et al., 2019), strengths of the novel sensor system are introduced to improve the current environment perception system. For now, only a qualitative observation of the capability to detect lane markings and other landmarks can be presented. Future work will use the current findings for landmark detection for a vehicle self-localization system.

Detection of Landmarks by Autonomous Mobile Robots Using Camera-Based Sensors in Outdoor Environments

This paper presents the methods of increasing the probability of landmark detection using the features of the color parameters (coordinates) of the image averaged over each column of the camera matrix pixels. The first method involves selecting a jump (dip) of color parameters distribution and averaging these parameters in the background areas using shifted similar distributions for images obtained by scanning the surrounding space. This possibility occurs if the correlation radius of the color parameters on the image is several times smaller than the horizontal size of the image. The second method uses a color parameter jump on the first modes of the Hilbert-Huang transform in the location of a landmark image.