Electronic portal image enhancement based on nonuniformity correction in wavelet domain.

PURPOSE: Electronic portal images are one of the most important tools to verify the ongoing radiotherapy treatment through comparison with a reference image generated during treatment planning. In this procedure, two images are geometrically matched by means of visible bone or other landmarks of interest such as implanted fiducials. However, the intrinsically poor contrast and low spatial resolution of portal images can limit image quality. METHODS: In this study, we have provided a multiresolution approach to enhance the quality of portal images acquired from the pelvis treatment fields. The main idea behind this work aims at removing some of the image artifacts that conceal the anatomical information. For this purpose, we have applied the homomorphic filtering on the approximation sub-band of wavelet decomposition to enhance local information. Moreover, in order to sharpen the bone edges, wavelet detail sub-bands were weighted to amplify important image details in the reconstruction of the desired enhanced image. The most appropriate image quality measure was chosen according to the image's characteristics in the spatial domain. By considering the characteristics of portal images as the random and nonperiodic texture, high level of noise, and a nonuniform background, three suitable quality measures of images were assessed: edge content, measure of enhancement, and measure of enhancement by entropy. RESULTS: The higher values of these measures indicate the quality improvement in the processed images through our proposed algorithm. Moreover, the subjective evaluation results indicate that the proposed multiresolution approach significantly enhances the perceived quality of images in comparison with original and the similar approach ( p < 0.001 $p < 0.001$ ). CONCLUSIONS: Our proposed wavelet-based enhancement algorithm successfully reduced image intensity nonuniformity and enhanced anatomical featured information, which drastically improved the objective metrics values. Subjective evaluation of enhanced image confirmed this quality improvement.

Development and feasibility of a soft pneumatic-robotic glove to assist impaired hand function in quadriplegia patients: A pilot study.

One of the common disorders in people with quadriplegia is having a weak grip strength that can affect activities of daily living (ADL). This study presents the design of a soft robotic glove via pneumatic actuators and feasibility according to a range of motion (ROM) of proximal interphalangeal (PIP) joint and user friendly. The soft robotic glove includes a neoprene cockup, two pumps, a controller that adjusts the pressure of the pumps, two-direction parts, and two silicone tubes placed on an index and middle fingers. A total of seven subjects (2healthy, 5quadriplegia patients) participated in this project. Performance of the device was verified through assessment in healthy participants first and then spinal cord injury (SCI) participants. The device evaluated the range of motion (ROM) of a proximal interphalangeal (PIP) joint. Then, subjects completed a satisfaction questionnaire. Results showed the ROM of the PIP joint (p value = 0.042) increased by using the robotic glove. The average score of the satisfaction questionnaire was 4.24 which was beyond the desirable threshold. In conclusion, the glove obtained ROM requirements to the grip usual objects and underlined the potential for assisting SCI participants in ADLs. Providing motion in all fingers should be investigated and developed in the future.

Neutron spectrum unfolding using radial basis function neural networks.

Neutron energy spectrum unfolding has been the subject of research for several years. The Bayesian theory, Monte Carlo simulation, and iterative methods are some of the methods that have been used for neutron spectrum unfolding. In this study, the radial basis function (RBF), multilayer perceptron, and artificial neural networks (ANNs) were used for the unfolding of neutron spectrum, and a comparison was made between the networks' results. Both neural network architectures were trained and tested using the same data set for neutron spectrum unfolding from the response of LiI detectors with Eu impurity. Advantages of each ANN method in the unfolding of neutron energy spectrum were investigated, and the performance of the networks was compared. The results obtained showed that RBF neural network can be applied as an effective method for unfolding neutron spectrum, especially when the main target is the neutron dosimetry.