Vision-Based People Detection System for Heavy Machine Applications.

This paper presents a vision-based people detection system for improving safety in heavy machines. We propose a perception system composed of a monocular fisheye camera and a LiDAR. Fisheye cameras have the advantage of a wide field-of-view, but the strong distortions that they create must be handled at the detection stage. Since people detection in fisheye images has not been well studied, we focus on investigating and quantifying the impact that strong radial distortions have on the appearance of people, and we propose approaches for handling this specificity, adapted from state-of-the-art people detection approaches. These adaptive approaches nevertheless have the drawback of high computational cost and complexity. Consequently, we also present a framework for harnessing the LiDAR modality in order to enhance the detection algorithm for different camera positions. A sequential LiDAR-based fusion architecture is used, which addresses directly the problem of reducing false detections and computational cost in an exclusively vision-based system. A heavy machine dataset was built, and different experiments were carried out to evaluate the performance of the system. The results are promising, in terms of both processing speed and performance.

Analytic calculations and synthetic shapes for validation of quantitative contour comparison software.

A high level of variability in reported values was observed in a recent survey of contour similarity measures (CSMs) calculation tools. Such variations in the output measurements prevent meaningful comparison between studies. The purpose of this study was to develop a dataset with analytically calculated gold standard values to facilitate standardization and ensure accuracy of CSM implementations. The dataset was generated in the Digital Imaging and Communications in Medicine (DICOM) format. Both the dataset and the software used for its generation are made publicly available to encourage robust testing of CSM implementations for accuracy, improving consistency between different implementations.

An approach to generate synthetic 4DCT datasets to benchmark Mid-Position implementations.

PURPOSE: The Mid-Position image is constructed from 4DCT data using Deformable Image Registration and can be used as planning CT with reduced PTV volumes. 4DCT datasets currently-available for testing do not provide the corresponding Mid-P images of the datasets. This work describes an approach to generate human-like synthetic 4DCT datasets with the associated Mid-P images that can be used as reference in the validation of Mid-P implementations. METHODS: Twenty synthetic 4DCT datasets with the associated reference Mid-P images were generated from twenty clinical 4DCT datasets. Per clinical dataset, an anchor phase was registered to the remaining nine phases to obtain nine Deformable Vector Fields (DVFs). These DVFs were used to warp the anchor phase in order to generate the synthetic 4DCT dataset and the corresponding reference Mid-P image. Similarly, a reference 4D tumor mask dataset and its corresponding Mid-P tumor mask were generated. The generated synthetic datasets and masks were used to compare and benchmark the outcomes of three independent Mid-P implementations using a set of experiments. RESULTS: The Mid-P images constructed by the three implementations showed high similarity scores when compared to the reference Mid-P images except for one noisy dataset. The biggest difference in the estimated motion amplitudes (-2.6 mm) was noticed in the Superior-Inferior direction. The statistical analysis showed no significant differences among the three implementations for all experiments. CONCLUSION: The described approach and the proposed experiments provide an independent method that can be used in the validation of any Mid-P implementation being developed.

Multicenter comparison of measures for quantitative evaluation of contouring in radiotherapy.

Background and Purpose: A wide range of quantitative measures are available to facilitate clinical implementation of auto-contouring software, on-going Quality Assurance (QA) and interobserver contouring variation studies. This study aimed to assess the variation in output when applying different implementations of the measures to the same data in order to investigate how consistently such measures are defined and implemented in radiation oncology. Materials and Methods: A survey was conducted to assess if there were any differences in definitions of contouring measures or their implementations that would lead to variation in reported results between institutions. This took two forms: a set of computed tomography (CT) image data with "Test" and "Reference" contours was distributed for participants to process using their preferred tools and report results, and a questionnaire regarding the definition of measures and their implementation was completed by the participants. Results: Thirteen participants completed the survey and submitted results, with one commercial and twelve in-house solutions represented. Excluding outliers, variations of up to 50% in Dice Similarity Coefficient (DSC), 50% in 3D Hausdorff Distance (HD), and 200% in Average Distance (AD) were observed between the participant submitted results. Collaborative investigation with participants revealed a large number of bugs in implementation, confounding the understanding of intentional implementation choices. Conclusion: Care must be taken when comparing quantitative results between different studies. There is a need for a dataset with clearly defined measures and ground truth for validation of such tools prior to their use.

Real-world analysis of manual editing of deep learning contouring in the thorax region.

Background and purpose: User-adjustments after deep-learning (DL) contouring in radiotherapy were evaluated to get insight in real-world editing during clinical practice. This study assessed the amount, type and spatial regions of editing of auto-contouring for organs-at-risk (OARs) in routine clinical workflow for patients in the thorax region. Materials and methods: A total of 350 lung cancer and 362 breast cancer patients, contoured between March 2020 and March 2021 using a commercial DL-contouring method followed by manual adjustments were retrospectively analyzed. Subsampling was performed for some OARs, using an inter-slice gap of 1-3 slices. Commonly-used whole-organ contouring assessment measures were calculated, and all cases were registered to a common reference shape per OAR to identify regions of manual adjustment. Results were expressed as the median, 10th-90th percentile of adjustment and visualized using 3D renderings. Results: Per OAR, the median amount of editing was below 1 mm. However, large adjustments were found in some locations for most OARs. In general, enlarging of the auto-contours was needed. Subsampling DL-contours showed less adjustments were made in the interpolated slices compared to simulated no-subsampling for these OARs. Conclusion: The real-world performance of automatic DL-contouring software was evaluated and proven useful in clinical practice. Specific regions-of-adjustment were identified per OAR in the thorax region, and separate models were found to be necessary for specific clinical indications different from training data. This analysis showed the need to perform routine clinical analysis especially when procedures or acquisition protocols change to have the best configuration of the workflow.

Interactive contouring through contextual deep learning.

PURPOSE: To investigate a deep learning approach that enables three-dimensional (3D) segmentation of an arbitrary structure of interest given a user provided two-dimensional (2D) contour for context. Such an approach could decrease delineation times and improve contouring consistency, particularly for anatomical structures for which no automatic segmentation tools exist. METHODS: A series of deep learning segmentation models using a Recurrent Residual U-Net with attention gates was trained with a successively expanding training set. Contextual information was provided to the models, using a previously contoured slice as an input, in addition to the slice to be contoured. In total, 6 models were developed, and 19 different anatomical structures were used for training and testing. Each of the models was evaluated for all 19 structures, even if they were excluded from the training set, in order to assess the model's ability to segment unseen structures of interest. Each model's performance was evaluated using the Dice similarity coefficient (DSC), Hausdorff distance, and relative added path length (APL). RESULTS: The segmentation performance for seen and unseen structures improved when the training set was expanded by addition of structures previously excluded from the training set. A model trained exclusively on heart structures achieved a DSC of 0.33, HD of 44 mm, and relative APL of 0.85 when segmenting the spleen, whereas a model trained on a diverse set of structures, but still excluding the spleen, achieved a DSC of 0.80, HD of 13 mm, and relative APL of 0.35. Iterative prediction performed better compared to direct prediction when considering unseen structures. CONCLUSIONS: Training a contextual deep learning model on a diverse set of structures increases the segmentation performance for the structures in the training set, but importantly enables the model to generalize and make predictions even for unseen structures that were not represented in the training set. This shows that user-provided context can be incorporated into deep learning contouring to facilitate semi-automatic segmentation of CT images for any given structure. Such an approach can enable faster de-novo contouring in clinical practice.

Assessment of manual adjustment performed in clinical practice following deep learning contouring for head and neck organs at risk in radiotherapy.

BACKGROUND AND PURPOSE: Auto-contouring performance has been widely studied in development and commissioning studies in radiotherapy, and its impact on clinical workflow assessed in that context. This study aimed to evaluate the manual adjustment of auto-contouring in routine clinical practice and to identify improvements regarding the auto-contouring model and clinical user interaction, to improve the efficiency of auto-contouring. MATERIALS AND METHODS: A total of 103 clinical head and neck cancer cases, contoured using a commercial deep-learning contouring system and subsequently checked and edited for clinical use were retrospectively taken from clinical data over a twelve-month period (April 2019-April 2020). The amount of adjustment performed was calculated, and all cases were registered to a common reference frame for assessment purposes. The median, 10th and 90th percentile of adjustment were calculated and displayed using 3D renderings of structures to visually assess systematic and random adjustment. Results were also compared to inter-observer variation reported previously. Assessment was performed for both the whole structures and for regional sub-structures, and according to the radiation therapy technologist (RTT) who edited the contour. RESULTS: The median amount of adjustment was low for all structures (<2 mm), although large local adjustment was observed for some structures. The median was systematically greater or equal to zero, indicating that the auto-contouring tends to under-segment the desired contour. CONCLUSION: Auto-contouring performance assessment in routine clinical practice has identified systematic improvements required technically, but also highlighted the need for continued RTT training to ensure adherence to guidelines.

An Evaluation of Atlas Selection Methods for Atlas-Based Automatic Segmentation in Radiotherapy Treatment Planning.

Atlas-based automatic segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed as a way to improve the accuracy and execution time of segmentation, assuming that, the more similar the atlas is to the patient, the better the results will be. This paper presents an analysis of atlas selection methods in the context of radiotherapy treatment planning. For a range of commonly contoured OARs, a thorough comparison of a large class of typical atlas selection methods has been performed. For this evaluation, clinically contoured CT images of the head and neck ( N=316 ) and thorax ( N=280 ) were used. The state-of-the-art intensity and deformation similarity-based atlas selection methods were found to compare poorly to perfect atlas selection. Counter-intuitively, atlas selection methods based on a fixed set of representative atlases outperformed atlas selection methods based on the patient image. This study suggests that atlas-based segmentation with currently available selection methods compares poorly to the potential best performance, hampering the clinical utility of atlas-based segmentation. Effective atlas selection remains an open challenge in atlas-based segmentation for radiotherapy planning.

Can Atlas-Based Auto-Segmentation Ever Be Perfect? Insights From Extreme Value Theory.

Atlas-based segmentation is used in radiotherapy planning to accelerate the delineation of organs at risk (OARs). Atlas selection has been proposed to improve the performance of segmentation, assuming that the more similar the atlas is to the patient, the better the result. It follows that the larger the database of atlases from which to select, the better the results should be. This paper seeks to estimate a clinically achievable expected performance under this assumption. Assuming a perfect atlas selection, an extreme value theory has been applied to estimate the accuracy of single-atlas and multi-atlas segmentation given a large database of atlases. For this purpose, clinical contours of most common OARs on computed tomography of the head and neck ( N=316 ) and thoracic ( N=280 ) cases were used. This paper found that while for most organs, perfect segmentation cannot be reasonably expected, auto-contouring performance of a level corresponding to clinical quality could be consistently expected given a database of 5000 atlases under the assumption of perfect atlas selection.

Ultrasound image segmentation: a survey.

This paper reviews ultrasound segmentation paper methods, in a broad sense, focusing on techniques developed for medical B-mode ultrasound images. First, we present a review of articles by clinical application to highlight the approaches that have been investigated and degree of validation that has been done in different clinical domains. Then, we present a classification of methodology in terms of use of prior information. We conclude by selecting ten papers which have presented original ideas that have demonstrated particular clinical usefulness or potential specific to the ultrasound segmentation problem.

Phase-based level set segmentation of ultrasound images.

Ultrasonic image segmentation is a difficult problem due to speckle noise, low contrast, and local changes of intensity. Intensity-based methods do not perform particularly well on ultrasound images. However, it has been previously shown that these images respond well to local phase-based methods which are theoretically intensity invariant. Here, we use level set propagation to capture the left ventricle boundaries. The proposed approach uses a new speed term based on local phase and local orientation derived from the monogenic signal, which makes the algorithm robust to attenuation artifact. Furthermore, we use Cauchy kernels, as a better alternative to the commonly used log-Gabor, as pair of quadrature filters for the feature extraction. Results on synthetic and natural data show that the proposed method can robustly handle noise, and captures well the low contrast boundaries.

Fully automatic kidneys detection in 2D CT images: a statistical approach.

In this paper, we focus on automatic kidneys detection in 2D abdominal computed tomography (CT) images. Identifying abdominal organs is one of the essential steps for visualization and for providing assistance in teaching, clinical training and diagnosis. It is also a key step in medical image retrieval application. However, due to gray levels similarities of adjacent organs, contrast media effect and relatively high variation of organ's positions and shapes, automatically identifying abdominal organs has always been a challenging task. In this paper, we present an original method, in a statistical framework, for fully automatic kidneys detection. It makes use of spatial and gray-levels prior models built using a set of training images. The method is tested on over 400 clinically acquired images and very promising results are obtained.

Velocity estimation in ultrasound images: a block matching approach.

In this paper, we focus on velocity estimation in ultrasound images sequences. Ultrasound images present many difficulties in image processing because of the typically high level of noise found in them. Recently, Cohen and Dinstein have derived a new similarity measure, according to a simplified image formation model of ultrasound images, optimal in the maximum likelihood sense. This similarity measure is better for ultrasound images than others such as the sum-of-square differences or normalised cross-correlation because it takes into account the fact that the noise in an ultrasound image is multiplicative Rayleigh noise, and that displayed ultrasound images are log-compressed. In this work we investigate the use of this similarity measure in a block matching method. The underlying framework of the method is Singh's algorithm. New improvements are made both on the similarity measure and the Singh algorithm to provide better velocity estimates. A global optimisation scheme for algorithm parameter estimation is also proposed. We show that this optimisation makes an improvement of approximately 35% in comparison to the result obtained with the worst parameter set. Results on clinically acquired cardiac and breast ultrasound sequences, demonstrate the robustness of the method.