Acceleration of deep neural network (DNN) inference has gained increasing attention recently with the wide adoption of DNNs for practical applications. For computer vision tasks where inputs are images, existing works mostly focus on improving the throughput of inference for multiple images. However, in many real-time applications, it is critical to reduce the latency of a single image inference, which is more complicated than improving the throughput because of the inherent data dependencies. On the other hand, from human brain's perspective, the complexity in our visual surroundings is first encoded as a pattern of light on a two dimensional array of photoreceptors, with little direct resemblance to the original input or the ultimate percept. Within just a few hundred microns of retinal thickness, this initial signal encoded by our photoreceptors must be transformed into an adequate representation of the entire visual scene. Inspired by how the retina helps human brain incept new information efficiently, we present an end-to-end structured framework built using any existing convolutional neural network (CNN) as the backbone. The proposed framework, called VisualNet, can create task parallelism for the backbone during the inference of a single image. Experiments using a number of neural networks for the ImageNet classification task and the CIFAR-10 classification task on GPUs and CPUs show that the proposed VisualNet reduces the latency of the regular network it builds on by up to 80.6% when both are fully parallelized with state-of-the-art acceleration libraries. At the same time, VisualNet can achieve similar or slightly higher accuracy.

Total anomalous pulmonary venous connection (TAPVC) is a rare but mortal congenital heart disease in children and can be repaired by surgical operations. However, some patients may suffer from pulmonary venous obstruction (PVO) after surgery with insufficient blood supply, necessitating special follow-up strategy and treatment. Therefore, it is a clinically important yet challenging problem to predict such patients before surgery. In this paper, we address this issue and propose a computational framework to determine the risk factors for postoperative PVO (PPVO) from computed tomography angiography (CTA) images and build the PPVO risk prediction model. From clinical experiences, such risk factors are likely from the left atrium (LA) and pulmonary vein (PV) of the patient. Thus, 3D models of LA and PV are first reconstructed from low-dose CTA images. Then, a feature pool is built by computing different morphological features from 3D models of LA and PV, and the coupling spatial features of LA and PV. Finally, four risk factors are identified from the feature pool using the machine learning techniques, followed by a risk prediction model. As a result, not only PPVO patients can be effectively predicted but also qualitative risk factors reported in the literature can now be quantified. Finally, the risk prediction model is evaluated on two independent clinical datasets from two hospitals. The model can achieve the AUC values of 0.88 and 0.87 respectively, demonstrating its effectiveness in risk prediction.

Medical images may contain various types of artifacts with different patterns and mixtures, which depend on many factors such as scan setting, machine condition, patients’ characteristics, surrounding environment, etc. However, existing deep-learning-based artifact reduction methods are restricted by their training set with specific predetermined artifact types and patterns. As such, they have limited clinical adoption. In this paper, we introduce One-Shot medical image Artifact Reduction (OSAR), which exploits the power of deep learning but without using pre-trained general networks. Specifically, we train a light-weight image-specific artifact reduction network using data synthesized from the input image at test-time. Without requiring any prior large training data set, OSAR can work with almost any medical images that contain varying additive artifacts which are not in any existing data sets. In addition, Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) are used as vehicles and show that the proposed method can reduce artifacts better than state-of-the-art both qualitatively and quantitatively using shorter test time.

Automatic myocardial segmentation of contrast echocardio-graphy has shown great potential in the quantification of myocardial perfusion parameters. Segmentation quality control is an important step to ensure the accuracy of segmentation results for quality research as well as its clinical application. Usually, the segmentation quality control happens after the data acquisition. At the data acquisition time, the operator could not know the quality of the segmentation results. On-the-fly segmentation quality control could help the operator to adjust the ultrasound probe or retake data if the quality is unsatisfied, which can greatly reduce the effort of time-consuming manual correction. However, it is infeasible to deploy state-of-the-art DNN-based models because the segmentation module and quality control module must fit in the limited hardware resource on the ultrasound machine while satisfying strict latency constraints. In this paper, we propose a hardware-aware neural architecture search framework for automatic myocardial segmentation and quality control of contrast echocardiography. We explicitly incorporate the hardware latency as a regularization term into the loss function during training. The proposed method searches the best neural network architecture for the segmentation module and quality prediction module with strict latency.

Deep neural networks have demonstrated their great potential in recent years, exceeding the performance of human experts in a wide range of applications. Due to their large sizes, however, compression techniques such as weight quantization and pruning are usually applied before they can be accommodated on the edge. It is generally believed that quantization leads to performance degradation, and plenty of existing works have explored quantization strategies aiming at minimum accuracy loss. In this paper, we argue that quantization, which essentially imposes regularization on weight representations, can sometimes help to improve accuracy. We conduct comprehensive experiments on three widely used applications: fully connected network for biomedical image segmentation, convolutional neural network for image classification on ImageNet, and recurrent neural network for automatic speech recognition, and experimental results show that quantization can improve the accuracy by 1%, 1.95%, 4.23% on the three applications respectively with 3.5x-6.4x memory reduction.

Congenital heart disease (CHD) is the most common birth defect, accounting for one-third of all congenital birth defects. As with complicated intracardiac structural abnormalities, CHD is usually treated with surgical repair, and computed tomography (CT) is the main examination method for diagnosis of CHD and also provides anatomical information to surgeons. Currently, there exists a serious shortage of professional surgeons in developing countries. Compared with developed countries where large hospitals and cardiovascular disease centers have professional surgical teams with rich treatment experience, surgeons in developing countries and remote areas suffer from lack of professional surgical skills resulting with low surgical quality and high mortality. Recently, surgical telementoring has been popular to tackle the above problems, in which less-skilled surgeons can get real-time guidance from skilled surgeons remotely through audio and video transmission. However, there still exists difficulties in applying telementoring to CHD surgeries including high resource consumption on medical data transmission and storage, large image noise, and inconvenient and inefficient discussion between surgeons on CT. In this article, we proposed a framework with an image compression module, an image denoising module, and an image segmentation module based on CT images in CHD. We evaluated the above three modules and compared them with existing works, respectively, and the results show that our methods achieve much better performance. Furthermore, with 3D printing, VR technology, and 5G communications, our framework was successfully used in a real case study to treat a patient who needed surgical treatment.

As one of the most commonly ordered imaging tests, the computed tomography (CT) scan comes with inevitable radiation exposure that increases cancer risk to patients. However, CT image quality is directly related to radiation dose, and thus it is desirable to obtain high-quality CT images with as little dose as possible. CT image denoising tries to obtain high-dose-like high-quality CT images (domain Y ) from low dose low-quality CT images (domain X ), which can be treated as an image-to-image translation task where the goal is to learn the transform between a source domain X (noisy images) and a target domain Y (clean images). Recently, the cycle-consistent adversarial denoising network (CCADN) has achieved state-of-the-art results by enforcing cycle-consistent loss without the need of paired training data, since the paired data is hard to collect due to patients’ interests and cardiac motion. However, out of concerns on patients’ privacy and data security, protocols typically require clinics to perform medical image processing tasks including CT image denoising locally (i.e., edge denoising). Therefore, the network models need to achieve high performance under various computation resource constraints including memory and performance. Our detailed analysis of CCADN raises a number of interesting questions that point to potential ways to further improve its performance using the same or even fewer computation resources. For example, if the noise is large leading to a significant difference between domain X and domain Y , can we bridge X and Y with a intermediate domain Z such that both the denoising process between X and Z and that between Z and Y are easier to learn? As such intermediate domains lead to multiple cycles, how do we best enforce cycle- consistency? Driven by these questions, we propose a multi-cycle-consistent adversarial network (MCCAN) that builds intermediate domains and enforces both local and global cycle-consistency for edge denoising of CT images. The global cycle-consistency couples all generators together to model the whole denoising process, whereas the local cycle-consistency imposes effective supervision on the process between adjacent domains. Experiments show that both local and global cycle-consistency are important for the success of MCCAN, which outperforms CCADN in terms of denoising quality with slightly less computation resource consumption.

Deep learning had already demonstrated its power in medical images, including denoising, classification, segmentation, etc. All these applications are proposed to automatically analyze medical images beforehand, which brings more information to radiologists during clinical assessment for accuracy improvement. Recently, many medical denoising methods had shown their significant artifact reduction result and noise removal both quantitatively and qualitatively. However, those existing methods are developed around human-vision, i.e., they are designed to minimize the noise effect that can be perceived by human eyes. In this paper, we introduce an application-guided denoising framework, which focuses on denoising for the following neural networks. In our experiments, we apply the proposed framework to different datasets, models, and use cases. Experimental results show that our proposed framework can achieve a better result than human-vision denoising network.

Real-time cardiac magnetic resonance imaging (MRI) plays an increasingly important role in guiding various cardiac interventions. In order to provide better visual assistance, the cine MRI frames need to be segmented on-the-fly to avoid noticeable visual lag. In addition, considering reliability and patient data privacy, the computation is preferably done on local hardware. State-of-the-art MRI segmentation methods mostly focus on accuracy only, and can hardly be adopted for real-time application or on local hardware. In this work, we present the first hardware-aware multi-scale neural architecture search (NAS) framework for real-time 3D cardiac cine MRI segmentation. The proposed framework incorporates a latency regularization term into the loss function to handle realtime constraints, with the consideration of underlying hardware. In addition, the formulation is fully differentiable with respect to the architecture parameters, so that stochastic gradient descent (SGD) can be used for optimization to reduce the computation cost while maintaining optimization quality. Experimental results on ACDC MICCAI 2017 dataset demonstrate that our hardware-aware multi-scale NAS framework can reduce the latency by up to 3.5× and satisfy the real-time constraints, while still achieving competitive segmentation accuracy, compared with the state-of-the-art NAS segmentation framework.

Cloud based medical image analysis has become popular recently due to the high computation complexities of various deep neural network (DNN) based frameworks and the increasingly large volume of medical images that need to be processed. It has been demonstrated that for medical images the transmission from local to clouds is much more expensive than the computation in the clouds itself. Towards this, 3D image compression techniques have been widely applied to reduce the data traffic. However, most of the existing image compression techniques are developed around human vision, i.e., they are designed to minimize distortions that can be perceived by human eyes. In this paper, we will use deep learning based medical image segmentation as a vehicle and demonstrate that interestingly, machine and human view the compression quality differently. Medical images compressed with good quality w.r.t. human vision may result in inferior segmentation accuracy. We then design a machine vision oriented 3D image compression framework tailored for segmentation using DNNs. Our method automatically extracts and retains image features that are most important to the segmentation. Comprehensive experiments on widely adopted segmentation frameworks with HVSMR 2016 challenge dataset show that our method can achieve significantly higher segmentation accuracy at the same compression rate, or much better compression rate under the same segmentation accuracy, when compared with the existing JPEG 2000 method. To the best of the authors' knowledge, this is the first machine vision guided medical image compression framework for segmentation in the clouds.