Orthognathic surgery corrects jaw deformities to improve aesthetics and functions. Due to the complexity of the craniomaxillofacial (CMF) anatomy, orthognathic surgery requires precise surgical planning, which involves predicting postoperative changes in facial appearance. To this end, most conventional methods involve simulation with biomechanical modeling methods, which are labor intensive and computationally expensive. Here we introduce a learning-based framework to speed up the simulation of postoperative facial appearances. Specifically, we introduce a facial shape change prediction network (FSC-Net) to learn the nonlinear mapping from bony shape changes to facial shape changes. FSC-Net is a point transform network weakly-supervised by paired preoperative and postoperative data without point-wise correspondence. In FSC-Net, a distance-guided shape loss places more emphasis on the jaw region. A local point constraint loss restricts point displacements to preserve the topology and smoothness of the surface mesh after point transformation. Evaluation results indicate that FSC-Net achieves 15× speedup with accuracy comparable to a state-of-the-art (SOTA) finite-element modeling (FEM) method.

Federated learning is an emerging paradigm allowing large-scale decentralized learning without sharing data across different data owners, which helps address the concern of data privacy in medical image analysis. However, the requirement for label consistency across clients by the existing methods largely narrows its application scope. In practice, each clinical site may only annotate certain organs of interest with partial or no overlap with other sites. Incorporating such partially labeled data into a unified federation is an unexplored problem with clinical significance and urgency. This work tackles the challenge by using a novel federated multi-encoding U-Net (Fed-MENU) method for multi-organ segmentation. In our method, a multi-encoding U-Net (MENU-Net) is proposed to extract organ-specific features through different encoding sub-networks. Each sub-network can be seen as an expert of a specific organ and trained for that client. Moreover, to encourage the organ-specific features extracted by different sub-networks to be informative and distinctive, we regularize the training of the MENU-Net by designing an auxiliary generic decoder (AGD). Extensive experiments on six public abdominal CT datasets show that our Fed-MENU method can effectively obtain a federated learning model using the partially labeled datasets with superior performance to other models trained by either localized or centralized learning methods. Source code is publicly available at https://github.com/DIAL-RPI/Fed-MENU.

Domain adaptation techniques have been demonstrated to be effective in addressing label deficiency challenges in medical image segmentation. However, conventional domain adaptation based approaches often concentrate on matching global marginal distributions between different domains in a class-agnostic fashion. In this paper, we present a dual-attention domain-adaptative segmentation network (DADASeg-Net) for cross-modality medical image segmentation. The key contribution of DADASeg-Net is a novel dual adversarial attention mechanism, which regularizes the domain adaptation module with two attention maps respectively from the space and class perspectives. Specifically, the spatial attention map guides the domain adaptation module to focus on regions that are challenging to align in adaptation. The class attention map encourages the domain adaptation module to capture class-specific instead of class-agnostic knowledge for distribution alignment. DADASeg-Net shows superior performance in two challenging medical image segmentation tasks.

Cephalometric analysis relies on accurate detection of craniomaxillofacial (CMF) landmarks from cone-beam computed tomography (CBCT) images. However, due to the complexity of CMF bony structures, it is difficult to localize landmarks efficiently and accurately. In this paper, we propose a deep learning framework to tackle this challenge by jointly digitalizing 105 CMF landmarks on CBCT images. By explicitly learning the local geometrical relationships between the landmarks, our approach extends Mask R-CNN for end-to-end prediction of landmark locations. Specifically, we first apply a detection network on a down-sampled 3D image to leverage global contextual information to predict the approximate locations of the landmarks. We subsequently leverage local information provided by higher-resolution image patches to refine the landmark locations. On patients with varying non-syndromic jaw deformities, our method achieves an average detection accuracy of 1.38± 0.95mm, outperforming a related state-of-the-art method.

Orthognathic surgical outcomes rely heavily on the quality of surgical planning. Automatic estimation of a reference facial bone shape significantly reduces experience-dependent variability and improves planning accuracy and efficiency. We propose an end-to-end deep learning framework to estimate patient-specific reference bony shape models for patients with orthognathic deformities. Specifically, we apply a point-cloud network to learn a vertex-wise deformation field from a patient's deformed bony shape, represented as a point cloud. The estimated deformation field is then used to correct the deformed bony shape to output a patient-specific reference bony surface model. To train our network effectively, we introduce a simulation strategy to synthesize deformed bones from any given normal bone, producing a relatively large and diverse dataset of shapes for training. Our method was evaluated using both synthetic and real patient data. Experimental results show that our framework estimates realistic reference bony shape models for patients with varying deformities. The performance of our method is consistently better than an existing method and several deep point-cloud networks. Our end-to-end estimation framework based on geometric deep learning shows great potential for improving clinical workflows.

Automatic craniomaxillofacial (CMF) landmark localization from cone-beam computed tomography (CBCT) images is challenging, considering that 1) the number of landmarks in the images may change due to varying deformities and traumatic defects, and 2) the CBCT images used in clinical practice are typically large. In this paper, we propose a two-stage, coarse-to-fine deep learning method to tackle these challenges with both speed and accuracy in mind. Specifically, we first use a 3D faster R-CNN to roughly locate landmarks in down-sampled CBCT images that have varying numbers of landmarks. By converting the landmark point detection problem to a generic object detection problem, our 3D faster R-CNN is formulated to detect virtual, fixed-size objects in small boxes with centers indicating the approximate locations of the landmarks. Based on the rough landmark locations, we then crop 3D patches from the high-resolution images and send them to a multi-scale UNet for the regression of heatmaps, from which the refined landmark locations are finally derived. We evaluated the proposed approach by detecting up to 18 landmarks on a real clinical dataset of CMF CBCT images with various conditions. Experiments show that our approach achieves state-of-the-art accuracy of 0.89 ± 0.64mm in an average time of 26.2 seconds per volume.

An increasing number of studies are leveraging unsupervised cross-modality synthesis to mitigate the limited label problem in training medical image segmentation models. They typically transfer ground truth annotations from a label-rich imaging modality to a label-lacking imaging modality, under an assumption that different modalities share the same anatomical structure information. However, since these methods commonly use voxel/pixel-wise cycle-consistency to regularize the mappings between modalities, high-level semantic information is not necessarily preserved. In this paper, we propose a novel anatomy-regularized representation learning approach for segmentation-oriented cross-modality image synthesis. It learns a common feature encoding across different modalities to form a shared latent space, where 1) the input and its synthesis present consistent anatomical structure information, and 2) the transformation between two images in one domain is preserved by their syntheses in another domain. We applied our method to the tasks of cross-modality skull segmentation and cardiac substructure segmentation. Experimental results demonstrate the superiority of our method in comparison with state-of-the-art cross-modality medical image segmentation methods.

Compared to computed tomography (CT), magnetic resonance imaging (MRI) delineation of craniomaxillofacial (CMF) bony structures can avoid harmful radiation exposure. However, bony boundaries are blurry in MRI, and structural information needs to be borrowed from CT during the training. This is challenging since paired MRI-CT data are typically scarce. In this paper, we propose to make full use of unpaired data, which are typically abundant, along with a single paired MRI-CT data to construct a one-shot generative adversarial model for automated MRI segmentation of CMF bony structures. Our model consists of a cross-modality image synthesis sub-network, which learns the mapping between CT and MRI, and an MRI segmentation sub-network. These two sub-networks are trained jointly in an end-to-end manner. Moreover, in the training phase, a neighbor-based anchoring method is proposed to reduce the ambiguity problem inherent in cross-modality synthesis, and a feature-matching-based semantic consistency constraint is proposed to encourage segmentation-oriented MRI synthesis. Experimental results demonstrate the superiority of our method both qualitatively and quantitatively in comparison with the state-of-the-art MRI segmentation methods.

It has been shown that it is impossible to achieve stringent timely delivery guarantees in a large network without having complete information of all future packet arrivals. In order to maintain desirable performance in the presence of uncertainty of future, a viable approach is to add redundancy by increasing link capacities. This paper studies the amount of capacity needed to provide stringent timely delivery guarantees. We propose a low-complexity online algorithm and prove that it only requires a small amount of redundancy to guarantee the timely delivery of most packets. Furthermore, we show that in large networks with very high timely delivery requirements, the redundancy needed by our policy is at most twice as large as the theoretical lower bound. For practical implementation, we propose a distributed protocol based on this centralized policy. Without adding redundancy, we further propose a low-complexity order-optimal online policy for the network. The simulation results show that our policies achieve much better performance than the other state-of-the-art policies.

WiFi offloading, where mobile users opportunistically obtain data through WiFi rather than cellular networks, is a promising technique for greatly improving spectrum efficiency and reduce cellular network congestion. We consider a system where the service provider deploys multiple WiFi hotspots to offload mobile traffic, and study the scheduling policy to maximize the amount of offloaded data. Since users' movements are unpredictable, we focus on online scheduling policy, where APs have no knowledge of users' mobility patterns. We study the performance of online policies by comparing them against the optimal offline policy. We prove that any work-conserving policy is able to offload at least half as much data as the offline policy, and then propose an online policy such that when the requested data by each user is very large, the policy can offload $(e-1)/e$ as much data as the offline policy, where $e$ is Euler's constant. We further study the case where the service provider can increase the capacity of WiFi so as to provide some guarantees on the amount of offloaded data. We derive a lower-bound on the trade-off between capacity and the amount of offloaded data, and propose a simple online policy that achieves this lower bound. In addition, we show that our policy only needs half as much capacity as current mechanisms to provide the same performance guarantee.

This paper studies the problem of using minimum power to provide satisfactory performance for real-time applications over unreliable and fading wireless channels. We demonstrate that this problem can be formulated as a linear programming problem. However, this formulation involves exponentially many constraints, and many parameters are either unavailable or difficult to compute, which makes it infeasible to employ standard techniques to solve the linear programming problem. Instead, we propose a simple online algorithm for this problem. We prove that our algorithm provides satisfactory performance to all real-time applications, and the total power consumption can be made arbitrarily close to the theoretical lower bound. Furthermore, our algorithm has very low complexity and does not require knowledge of many parameters in the linear programming problem, including the distributions of channel qualities. We further extend our algorithm to address systems where real-time applications and non-real-time ones coexist. We demonstrate that our algorithm achieves both low total power consumption and high utility for each non-real-time client while satisfying the performance requirements of real-time clients. Simulation results further provide some insights in setting important parameters of our algorithms, and demonstrate that our algorithm indeed achieves a significant reduction in power consumption.

The problem of allocating jobs to appropriate servers in cloud computing is studied in this paper. We consider that jobs of various types arrive in some unpredictable pattern and the system is required to allocate a certain ratio of jobs. In order to meet the hard allocation ratio requirement in the presence of unknown arrival patterns, one can increase the capacity of servers by expanding the size of data centers. We then aim to find the minimum capacity needed to meet a given allocation ratio requirement. We propose two online job allocation policies with low complexity. We prove that, given a hard allocation ratio requirement, these two policies can achieve the requirement with the least capacity. We also derive a closed-from expression for the amount of capacity needed to achieve any given requirement. Two other popular policies are studied, and we demonstrate that they need at least an order higher capacity to meet the same hard allocation ratio requirement. Simulation results demonstrate that our policies remain far superior than the other two when jobs arrive according to some random process.

WiFi offloading, where mobile users opportunistically obtain data through WiFi rather than through cellular networks, is a promising technique to greatly improve spectrum efficiency and reduce cellular network congestion. We consider a system where the service provider deploys multiple WiFi hotspots to offload mobile traffic, and study the scheduling policy to maximize the amount of offloaded data. Since the movements of users are unpredictable, we focus on online scheduling policy where APs do not have any knowledge about the users' mobility patterns. We study performance of online policies by comparing against the optimal offline policy. We prove that any work-conserving policy is able to offload at least half as much data as the offline policy, and then propose an online policy that can offload (e-1)/e as much data as the offline policy. We further study the case where the service provider can increase the capacity of WiFi so as to provide guarantees on the amount of offloaded data. We propose a simple online policy and prove that our policy only needs half as much capacity as current mechanism to provide the same performance guarantee.