University of Notre Dame Du Lac

University System of Maryland

The University of New Mexico

University of Iowa Research Foundation

The Chinese University of Hong Kong

The Hong Kong Polytechnic University

Danny Ziyi Chen

Cedric Xinsheng Yu

Shuang Luan

Xiaobo Sharon Hu

Chao Wang

Xiaodong Wu

Matthew A Earl

Milan Sonka

Xiaoming Chen

Jiazhuo Wang

John DeWolfe MacKenzie

Kang Li

Yizhe Zhang

Anil Tejbhan Ahuja

Rageshree Ramachandran

Michael Tin-Cheung Ying

Lin Yang

Bo Zhou

Kevin Whitton

A method operating a Graphics Processing Unit (GPU) memory can be provided by accessing specified training parameters used to train a Deep Neural Network (DNN) using a GPU with a local GPU memory, the specified training parameters including at least a specified batch size of samples configured to train the DNN. A sub-batch size of the samples can be defined that is less than or equal to the specified batch size of samples in response to determining that an available size of the local GPU memory is insufficient to store all data associated with training the DNN using one batch of the samples. Instructions configured to train the DNN using the sub-batch size can be defined so that an accuracy of the DNN trained using the sub-batch size is about equal to an accuracy of the DNN trained using the specified batch size of the samples.

Methods of operating a graphics processing unit (GPU) to train a deep neural network using a GPU local memory and related articles of manufacture

Optimizations are provided for segmenting tissue objects included in an ultrasound image. Initially, raw pixel data is received. Here, each pixel corresponds to ultrasound information. This raw pixel data is processed through a first fully convolutional network to generate a first segmentation label map. This first map includes a first set of objects that have been segmented into a coarse segmentation class. Then, this first map is processed through a second fully convolutional network to generate a second segmentation label map. This second map is processed using the raw pixel data as a base reference. Further, this second map includes a second set of objects that have been segmented into a fine segmentation class. Then, a contour optimization algorithm is applied to at least one of the second set of objects in order to refine that object's contour boundary. Subsequently, that object is identified as corresponding to a lymph node.

Segmenting ultrasound images

Systems and methods are provided for identifying markers for inflammation in a tissue image. The tissue image is captured as an image of a histology slide. Subcellular structures in the tissue image are segmented via a first automated process to identify at least one variety of immune cells within the image. Glands and vilii are identified within the tissue image via a second automated process. Neutrophils are identified within the tissue image via a third automated process. An output representing the identified glands, villi, neutrophils, and other immune cells is provided to a human operator.

Identification of inflammation in tissue images

Provided herein are methods and systems for designing a radiation treatment for a subject using single arc dose painting. The methods and systems comprise an algorithm or a computer-readable product having the same, to plan the radiation treatment. The algorithm converts pairs of multiple leaf collimation (MLC) leaves to sets of leaf aperture sequences that form a shortest path single arc thereof where the pairs of MLC leaves each aligned to an intensity profile of densely-spaced radiation beams, and connects each single arc of leaf apertures to form a final treatment single arc. Also provided is a method for irradiating a tumor in a subject using single arc dose painting.

Single-arc dose painting for precision radiation therapy

Disclosed is an example method to calculate radiation dose. The method includes receiving a tissue matrix in which the tissue matrix includes a plurality of voxels. The example method also includes producing a first plurality of transport lines with a direction controller in which each transport line is indicative of a cone of irradiated energy, and calculating at least one radiation dose with at least one deposit engine substantially in parallel with producing a second plurality of transport lines with the direction controller.

Methods and apparatus for hardware based radiation dose calculation

A system and methods for the efficient segmentation of globally optimal surfaces representing object boundaries in volumetric datasets is provided. An optical surface detection system and methods are provided that are capable of simultaneously detecting multiple interacting surfaces in which the optimality is controlled by the cost functions designed for individual surfaces and by several geometric constraints defining the surface smoothness and interrelations.

System and methods for image segmentation in n-dimensional space

The present invention provides a method comprising the following steps: (a) partitioning an intensity modulated beam into a set of sub-IMBs; and (b) partitioning the sub-IMBs into segments, wherein steps (a) and (b) introduce no machine delivery error. The present invention also provides a method comprising the following steps: (a) recursively partitioning an intensity modulated beam into plateaus; and (b) partitioning the plateaus into segments, wherein step (a) comprises determining a tradeoff between machine delivery error and the number of segments into which the plateaus will be partitioned in step (b).

Error control in algorithmic approach to step-and-shoot intensity modulated radiation therapy

The present invention provides a method comprising recursively partitioning an intensity modulated beam into plateaus; and partitioning the plateaus into segments. The present invention also provides a method for controlling administration of radiation therapy to a patient based on static leaf prescriptions. In addition, the present invention provides a method for partitioning an IMB.

Segmentation algorithmic approach to step-and-shoot intensity modulated radiation therapy

Growing community of inventors

Granger, IN, United States of America

Danny Ziyi Chen