Evaluation of blurring and noise of a medical image using a digital phantom

Blurring and noise are an essential characteristic of a medical image on the imaging system. This study shows the characteristics of blurring and noise of a medical image using a digital phantom. A square-shaped digital phantom was produced with pixels that consist of black and white. The line profile was selected on a binary digital image. An image with noise added was generated and a corresponding line profile was also drawn. The degree of noise was increased using the gaussian noise value. The blurring images obtained by applying gaussian blur to a digital phantom was produced similarities to real images. Also, the degree of blurring was increased using the gaussian blur value. As noise increased, the standard deviation of pixels inside and background the object also increased. However, the boundary of the object was retained. As image blurring increased, the boundary of the object was not clearly distinguished. However, the standard deviation of pixels inside and background the object was retained. When extreme noise and blurring are increased, the resulting images are different. For adding noise, the shape is visually maintained. However, the blurred image does not maintain a square shape. Therefore, it is shown that blurring due to movement of object cannot maintain original form. In the image processing method, the reduction of noise is achieved through blur processing. The noise was reduced through blur processing in the image with noise. The degree of noise decreased, but the ambiguity of the boundary increased.

A novel algorithm of the digital nervous tissue phantom creation based on 3D Voronoi diagram application

The essential part of mathematical modelling of nutrients convectional reaction-diffusion is creation of a digital phantom of considered biological object. This process becomes an especial problem which needs to be solved before numerical calculations of the concentration gradients will be done. There are two principal ways to get the solution in this case. The first approach is the reconstruction of a digital phantom on the base of the experimental data directly. The second one is the creation of a virtual object according to the experimental evidence and the known principals de novo. The main advantage of the created phantom is a high adaptability to modelling demands and a physical problem formulation. In the present study a new algorithm of a digital phantom creation has been established. The principles of the claimed procedures are demonstrated by the example of a nervous tissue. Initially, one needs to create N 3D objects according to Voronoi diagrams. Each object has 144 edges and 69 boundaries on average. Having chosen M rear objects, a long 3D structure mimicking neurons axons is created according to a loft procedure from the start boundaries to the end ones. Then, the set of Boolean operations has been applied to form continuous smooth objects. The remain (N-(M+s)) objects are combined into several whole bodies using the loft procedures between the closet neighbours. The final structure has a good conformity with a nervous tissue architecture. Furthermore, the obtained phantom is correct to the mesh application and further numerical calculations.

Measurement of Slice Thickness of a Medical Image using a Wedge Phantom

Slice thickness measurement is an essential parameter of performance evaluation for the medical imaging system. This study demonstrates the characteristics of slice thickness measurement for medical images using a wedge digital phantom. A wedge-shaped digital phantom was generated and the ideal edge response function (ERF) was extracted from line profile in single slice. The corresponding slice profile was calculated by the derivative of ERF. The wedge phantom obtained by applying gaussian convolving to a digital phantom was also generated to produce similarities to real medical images. Unlike an ideal slice profile, it was estimated by the full width half maximum (FWHM) of the Gaussian function fitting. In addition, we evaluate the effect of background noise and wedge angle for the wedge phantom. The estimated FWHM of the image with noise added was increased by 10.4% compared to the image without noise. However, the FWHM from the line profiles averaging on the noise-added image was estimated by 0.2% reduction than the noise-free image. The line profiles averaging improves the accurate measurement of slice thickness by decreasing the noise. Despite the wedge angle changing from 45 to 30 degrees, the resulting FWHM was estimated to have less than 1% difference. However, the length of the line profile to be acquired should be increased as the wedge angle increases.

Real-time markerless tumour tracking with patient-specific deep learning using a personalised data generation strategy: proof of concept by phantom study

Objective: For real-time markerless tumour tracking in stereotactic lung radiotherapy, we propose a different approach which uses patient-specific deep learning (DL) using a personalised data generation strategy, avoiding the need for collection of a large patient data set. We validated our strategy with digital phantom simulation and epoxy phantom studies. Methods: We developed lung tumour tracking for radiotherapy using a convolutional neural network trained for each phantom’s lesion by using multiple digitally reconstructed radiographs (DRRs) generated from each phantom’s treatment planning four-dimensional CT. We trained tumour-bone differentiation using large numbers of training DRRs generated with various projection geometries to simulate tumour motion. We solved the problem of using DRRs for training and X-ray images for tracking using the training DRRs with random contrast transformation and random noise addition. Results: We defined adequate tracking accuracy as the percentage frames satisfying <1 mm tracking error of the isocentre. In the simulation study, we achieved 100% tracking accuracy in 3 cm spherical and 1.5×2.25×3 cm ovoid masses. In the phantom study, we achieved 100 and 94.7% tracking accuracy in 3 cm and 2 cm spherical masses, respectively. This required 32.5 ms/frame (30.8 fps) real-time processing. Conclusions: We proved the potential feasibility of a real-time markerless tumour tracking framework for stereotactic lung radiotherapy based on patient-specific DL with personalised data generation with digital phantom and epoxy phantom studies. Advances in knowledge: Using DL with personalised data generation is an efficient strategy for real-time lung tumour tracking.