Dynamic Programming Based Segmentation in Biomedical Imaging

Combining natural and artificial intelligence for robust automatic anatomy segmentation: Application in neck and thorax auto-contouring

BACKGROUND

Automatic segmentation of 3D objects in computed tomography (CT) is challenging. Current methods, based mainly on artificial intelligence (AI) and end-to-end deep learning (DL) networks, are weak in garnering high-level anatomic information, which leads to compromised efficiency and robustness. This can be overcome by incorporating natural intelligence (NI) into AI methods via computational models of human anatomic knowledge.

PURPOSE

We formulate a hybrid intelligence (HI) approach that integrates the complementary strengths of NI and AI for organ segmentation in CT images and illustrate performance in the application of radiation therapy (RT) planning via multisite clinical evaluation.

METHODS

The system employs five modules: (i) body region recognition, which automatically trims a given image to a precisely defined target body region; (ii) NI-based automatic anatomy recognition object recognition (AAR-R), which performs object recognition in the trimmed image without DL and outputs a localized fuzzy model for each object; (iii) DL-based recognition (DL-R), which refines the coarse recognition results of AAR-R and outputs a stack of 2D bounding boxes (BBs) for each object; (iv) model morphing (MM), which deforms the AAR-R fuzzy model of each object guided by the BBs output by DL-R; and (v) DL-based delineation (DL-D), which employs the object containment information provided by MM to delineate each object. NI from (ii), AI from (i), (iii), and (v), and their combination from (iv) facilitate the HI system.

RESULTS

The HI system was tested on 26 organs in neck and thorax body regions on CT images obtained prospectively from 464 patients in a study involving four RT centers. Data sets from one separate independent institution involving 125 patients were employed in training/model building for each of the two body regions, whereas 104 and 110 data sets from the 4 RT centers were utilized for testing on neck and thorax, respectively. In the testing data sets, 83% of the images had limitations such as streak artifacts, poor contrast, shape distortion, pathology, or implants. The contours output by the HI system were compared to contours drawn in clinical practice at the four RT centers by utilizing an independently established ground-truth set of contours as reference. Three sets of measures were employed: accuracy via Dice coefficient (DC) and Hausdorff boundary distance (HD), subjective clinical acceptability via a blinded reader study, and efficiency by measuring human time saved in contouring by the HI system. Overall, the HI system achieved a mean DC of 0.78 and 0.87 and a mean HD of 2.22 and 4.53 mm for neck and thorax, respectively. It significantly outperformed clinical contouring in accuracy and saved overall 70% of human time over clinical contouring time, whereas acceptability scores varied significantly from site to site for both auto-contours and clinically drawn contours.

CONCLUSIONS

The HI system is observed to behave like an expert human in robustness in the contouring task but vastly more efficiently. It seems to use NI help where image information alone will not suffice to decide, first for the correct localization of the object and then for the precise delineation of the boundary.

A Study on Weak Edge Detection of COVID-19's CT Images Based on Histogram Equalization and Improved Canny Algorithm

The coronavirus disease 2019 (COVID-19) is a substantial threat to people's lives and health due to its high infectivity and rapid spread. Computed tomography (CT) scan is one of the important auxiliary methods for the clinical diagnosis of COVID-19. However, CT image lesion edge is normally affected by pixels with uneven grayscale and isolated noise, which makes weak edge detection of the COVID-19 lesion more complicated. In order to solve this problem, an edge detection method is proposed, which combines the histogram equalization and the improved Canny algorithm. Specifically, the histogram equalization is applied to enhance image contrast. In the improved Canny algorithm, the median filter, instead of the Gaussian filter, is used to remove the isolated noise points. The K-means algorithm is applied to separate the image background and edge. And the Canny algorithm is improved continuously by combining the mathematical morphology and the maximum between class variance method (OTSU). On selecting four types of lesion images from COVID-CT date set, MSE, MAE, SNR, and the running time are applied to evaluate the performance of the proposed method. The average values of these evaluation indicators are 1.7322, 7.9010, 57.1241, and 5.4887, respectively. Compared with other three methods, these values indicate that the proposed method achieves better result. The experimental results prove that the proposed algorithm can effectively detect the weak edge of the lesion, which is helpful for the diagnosis of COVID-19.

[Extraction of respiratory signals from chest tomosynthesis]

OBJECTIVE

To analyze the respiratory motion of the scanned object during acquisition of digital chest tomosynthesis (CTS) using a linear model.

OBJECTIVE

Respiratory signals were generated by extracting the motion of the diaphragm from the projection radiographs. The diaphragm trajectory obtained by dynamic programming (DP) was modeled and fitted, and according to the fitting of the data, the base motion curve and respiratory signal curve of the diaphragm were separated. Multipurpose chest phantom data, simulated digital Xcat phantom data and the datasets of 3 clinical patients were used to validate the performance of the proposed method.

OBJECTIVE

The motion trajectory of the diaphragm extracted from multipurpose chest phantom simulation data was linear. The respiratory signals could be effectively extracted from the 3 datasets of clinical patients in different respiratory states. The correlation coefficient between the respiratory signal extracted in Xcat simulation experiment and the original design was 0.9797.

OBJECTIVE

The linear model can effectively obtain the respiratory motion information of patients in real time, thus enabling the physicians to make clinical decisions on a rescan.

Skin Surface Detection in 3D Optoacoustic Mesoscopy Based on Dynamic Programming

Optoacoustic (photoacoustic) mesoscopy offers unique capabilities in skin imaging and resolves skin features associated with detection, diagnosis, and management of disease. A critical first step in the quantitative analysis of clinical optoacoustic images is to identify the skin surface in a rapid, reliable, and automated manner. Nevertheless, most common edge- and surface-detection algorithms cannot reliably detect the skin surface on 3D raster-scan optoacoustic mesoscopy (RSOM) images, due to discontinuities and diffuse interfaces in the image. We present herein a novel dynamic programming approach that extracts the skin boundary as a 2D surface in one single step, as opposed to consecutive extraction of several independent 1D contours. A domain-specific energy function is introduced, taking into account the properties of volumetric optoacoustic mesoscopy images. The accuracy of the proposed method is validated on scans of the volar forearm of 19 volunteers with different skin complexions, for which the skin surface has been traced manually to provide a reference. In addition, the robustness and the limitations of the method are demonstrated on data where the skin boundaries are low-contrast or ill-defined. The automatic skin surface detection method can improve the speed and accuracy in the analysis of quantitative features seen on the RSOM images and accelerate the clinical translation of the technique. Our method can likely be extended to identify other types of surfaces in the RSOM and other imaging modalities.

Sin-quadratic model for chest tomosynthesis respiratory signal analysis and its application in four dimensional chest tomosynthesis reconstruction

Chest tomosynthesis (CTS) is a newly developed imaging technique which provides pseudo-3D volume anatomical information of thorax from limited-angle projections and contains much less of superimposed anatomy than the chest X-ray radiography. One of the relatively common problems in CTS is the patient respiratory motion during image acquisition, which negatively impacts the detectability. In this work, we propose a sin-quadratic model to analyze the respiratory motion during CTS scan, which is a real time method where the respiratory signal is generated by extracting the motion of diaphragm from projection radiographs. According to the estimated respiratory signal, the CTS projections were then amplitude-based sorted into four to eight phases, and an iterative reconstruction strategy with total variation regularization was adopted to reconstruct the CTS images at each phase. Simulated digital XCAT phantom data and three sets of patient data were adopted for the experiments to validate the performance of the sin-quadratic model and its application in four dimensional (4D) CTS reconstruction. Results of the XCAT phantom simulation study show that the correlation coefficient between the extracted respiratory signal and the originally designed respiratory signal is 0.9964, which suggests that the proposed model could exactly extract the respiratory signal from CTS projections. The 4D CTS reconstructions of both the phantom data and the patient data show clear reduction of motion-induced blur.