Youbao Tang

Xiaoye Qian, Youbao Tang, Ning Zhang et al.

Transformer-based approaches have been successfully proposed for 3D human pose estimation (HPE) from 2D pose sequence and achieved state-of-the-art (SOTA) performance. However, current SOTAs have difficulties in modeling spatial-temporal correlations of joints at different levels simultaneously. This is due to the poses' spatial-temporal complexity. Poses move at various speeds temporarily with various joints and body-parts movement spatially. Hence, a cookie-cutter transformer is non-adaptable and can hardly meet the "in-the-wild" requirement. To mitigate this issue, we propose Hierarchical Spatial-Temporal transFormers (HSTFormer) to capture multi-level joints' spatial-temporal correlations from local to global gradually for accurate 3D HPE. HSTFormer consists of four transformer encoders (TEs) and a fusion module. To the best of our knowledge, HSTFormer is the first to study hierarchical TEs with multi-level fusion. Extensive experiments on three datasets (i.e., Human3.6M, MPI-INF-3DHP, and HumanEva) demonstrate that HSTFormer achieves competitive and consistent performance on benchmarks with various scales and difficulties. Specifically, it surpasses recent SOTAs on the challenging MPI-INF-3DHP dataset and small-scale HumanEva dataset, with a highly generalized systematic approach. The code is available at: https://github.com/qianxiaoye825/HSTFormer.

Youbao Tang, Ning Zhang, Yirui Wang et al.

Automatically measuring lesion/tumor size with RECIST (Response Evaluation Criteria In Solid Tumors) diameters and segmentation is important for computer-aided diagnosis. Although it has been studied in recent years, there is still space to improve its accuracy and robustness, such as (1) enhancing features by incorporating rich contextual information while keeping a high spatial resolution and (2) involving new tasks and losses for joint optimization. To reach this goal, this paper proposes a transformer-based network (MeaFormer, Measurement transFormer) for lesion RECIST diameter prediction and segmentation (LRDPS). It is formulated as three correlative and complementary tasks: lesion segmentation, heatmap prediction, and keypoint regression. To the best of our knowledge, it is the first time to use keypoint regression for RECIST diameter prediction. MeaFormer can enhance high-resolution features by employing transformers to capture their long-range dependencies. Two consistency losses are introduced to explicitly build relationships among these tasks for better optimization. Experiments show that MeaFormer achieves the state-of-the-art performance of LRDPS on the large-scale DeepLesion dataset and produces promising results of two downstream clinic-relevant tasks, i.e., 3D lesion segmentation and RECIST assessment in longitudinal studies.

Yirui Wang, Shenghua He, Youbao Tang et al.

In order to cope with the increasing demand for labeling data and privacy issues with human detection, synthetic data has been used as a substitute and showing promising results in human detection and tracking tasks. We participate in the 7th Workshop on Benchmarking Multi-Target Tracking (BMTT), themed on "How Far Can Synthetic Data Take us"? Our solution, PieTrack, is developed based on synthetic data without using any pre-trained weights. We propose a self-supervised domain adaptation method that enables mitigating the domain shift issue between the synthetic (e.g., MOTSynth) and real data (e.g., MOT17) without involving extra human labels. By leveraging the proposed multi-scale ensemble inference, we achieved a final HOTA score of 58.7 on the MOT17 testing set, ranked third place in the challenge.

Yifan Liu, Youbao Tang, Ning Zhang et al.

Temporal action localization (TAL) aims to detect the boundary and identify the class of each action instance in a long untrimmed video. Current approaches treat video frames homogeneously, and tend to give background and key objects excessive attention. This limits their sensitivity to localize action boundaries. To this end, we propose a prior-enhanced temporal action localization method (PETAL), which only takes in RGB input and incorporates action subjects as priors. This proposal leverages action subjects' information with a plug-and-play subject-aware spatial attention module (SA-SAM) to generate an aggregated and subject-prioritized representation. Experimental results on THUMOS-14 and ActivityNet-1.3 datasets demonstrate that the proposed PETAL achieves competitive performance using only RGB features, e.g., boosting mAP by 2.41% or 0.25% over the state-of-the-art approach that uses RGB features or with additional optical flow features on the THUMOS-14 dataset.

Zhangpu Li, Changhong Zou, Suxue Ma et al.

The rocketing prosperity of large language models (LLMs) in recent years has boosted the prevalence of vision-language models (VLMs) in the medical sector. In our online medical consultation scenario, a doctor responds to the texts and images provided by a patient in multiple rounds to diagnose her/his health condition, forming a multi-turn multimodal medical dialogue format. Unlike high-quality images captured by professional equipment in traditional medical visual question answering (Med-VQA), the images in our case are taken by patients' mobile phones. These images have poor quality control, with issues such as excessive background elements and the lesion area being significantly off-center, leading to degradation of vision-language alignment in the model training phase. In this paper, we propose ZALM3, a Zero-shot strategy to improve vision-language ALignment in Multi-turn Multimodal Medical dialogue. Since we observe that the preceding text conversations before an image can infer the regions of interest (RoIs) in the image, ZALM3 employs an LLM to summarize the keywords from the preceding context and a visual grounding model to extract the RoIs. The updated images eliminate unnecessary background noise and provide more effective vision-language alignment. To better evaluate our proposed method, we design a new subjective assessment metric for multi-turn unimodal/multimodal medical dialogue to provide a fine-grained performance comparison. Our experiments across three different clinical departments remarkably demonstrate the efficacy of ZALM3 with statistical significance.

Zhenxiao Liang, Ning Zhang, Youbao Tang et al.

We introduce ShapeGaussian, a high-fidelity, template-free method for 4D human reconstruction from casual monocular videos. Generic reconstruction methods lacking robust vision priors, such as 4DGS, struggle to capture high-deformation human motion without multi-view cues. While template-based approaches, primarily relying on SMPL, such as HUGS, can produce photorealistic results, they are highly susceptible to errors in human pose estimation, often leading to unrealistic artifacts. In contrast, ShapeGaussian effectively integrates template-free vision priors to achieve both high-fidelity and robust scene reconstructions. Our method follows a two-step pipeline: first, we learn a coarse, deformable geometry using pretrained models that estimate data-driven priors, providing a foundation for reconstruction. Then, we refine this geometry using a neural deformation model to capture fine-grained dynamic details. By leveraging 2D vision priors, we mitigate artifacts from erroneous pose estimation in template-based methods and employ multiple reference frames to resolve the invisibility issue of 2D keypoints in a template-free manner. Extensive experiments demonstrate that ShapeGaussian surpasses template-based methods in reconstruction accuracy, achieving superior visual quality and robustness across diverse human motions in casual monocular videos.

Ke Yan, Jinzheng Cai, Youjing Zheng et al.

Large-scale datasets with high-quality labels are desired for training accurate deep learning models. However, due to the annotation cost, datasets in medical imaging are often either partially-labeled or small. For example, DeepLesion is such a large-scale CT image dataset with lesions of various types, but it also has many unlabeled lesions (missing annotations). When training a lesion detector on a partially-labeled dataset, the missing annotations will generate incorrect negative signals and degrade the performance. Besides DeepLesion, there are several small single-type datasets, such as LUNA for lung nodules and LiTS for liver tumors. These datasets have heterogeneous label scopes, i.e., different lesion types are labeled in different datasets with other types ignored. In this work, we aim to develop a universal lesion detection algorithm to detect a variety of lesions. The problem of heterogeneous and partial labels is tackled. First, we build a simple yet effective lesion detection framework named Lesion ENSemble (LENS). LENS can efficiently learn from multiple heterogeneous lesion datasets in a multi-task fashion and leverage their synergy by proposal fusion. Next, we propose strategies to mine missing annotations from partially-labeled datasets by exploiting clinical prior knowledge and cross-dataset knowledge transfer. Finally, we train our framework on four public lesion datasets and evaluate it on 800 manually-labeled sub-volumes in DeepLesion. Our method brings a relative improvement of 49% compared to the current state-of-the-art approach in the metric of average sensitivity. We have publicly released our manual 3D annotations of DeepLesion in https://github.com/viggin/DeepLesion_manual_test_set.

Youbao Tang, Yuxing Tang, Jing Xiao et al.

This paper proposes a novel framework for lung segmentation in chest X-rays. It consists of two key contributions, a criss-cross attention based segmentation network and radiorealistic chest X-ray image synthesis (i.e. a synthesized radiograph that appears anatomically realistic) for data augmentation. The criss-cross attention modules capture rich global contextual information in both horizontal and vertical directions for all the pixels thus facilitating accurate lung segmentation. To reduce the manual annotation burden and to train a robust lung segmentor that can be adapted to pathological lungs with hazy lung boundaries, an image-to-image translation module is employed to synthesize radiorealistic abnormal CXRs from the source of normal ones for data augmentation. The lung masks of synthetic abnormal CXRs are propagated from the segmentation results of their normal counterparts, and then serve as pseudo masks for robust segmentor training. In addition, we annotate 100 CXRs with lung masks on a more challenging NIH Chest X-ray dataset containing both posterioranterior and anteroposterior views for evaluation. Extensive experiments validate the robustness and effectiveness of the proposed framework. The code and data can be found from https://github.com/rsummers11/CADLab/tree/master/Lung_Segmentation_XLSor .

Canyu Zhang, Youbao Tang, Ning Zhang et al.

Dance serves as a powerful medium for expressing human emotions, but the lifelike generation of dance is still a considerable challenge. Recently, diffusion models have showcased remarkable generative abilities across various domains. They hold promise for human motion generation due to their adaptable many-to-many nature. Nonetheless, current diffusion-based motion generation models often create entire motion sequences directly and unidirectionally, lacking focus on the motion with local and bidirectional enhancement. When choreographing high-quality dance movements, people need to take into account not only the musical context but also the nearby music-aligned dance motions. To authentically capture human behavior, we propose a Bidirectional Autoregressive Diffusion Model (BADM) for music-to-dance generation, where a bidirectional encoder is built to enforce that the generated dance is harmonious in both the forward and backward directions. To make the generated dance motion smoother, a local information decoder is built for local motion enhancement. The proposed framework is able to generate new motions based on the input conditions and nearby motions, which foresees individual motion slices iteratively and consolidates all predictions. To further refine the synchronicity between the generated dance and the beat, the beat information is incorporated as an input to generate better music-aligned dance movements. Experimental results demonstrate that the proposed model achieves state-of-the-art performance compared to existing unidirectional approaches on the prominent benchmark for music-to-dance generation.

Chi-Tung Cheng, Jinzheng Cai, Wei Teng et al.

Hepatocellular carcinoma (HCC) can be potentially discovered from abdominal computed tomography (CT) studies under varied clinical scenarios, e.g., fully dynamic contrast enhanced (DCE) studies, non-contrast (NC) plus venous phase (VP) abdominal studies, or NC-only studies. We develop a flexible three-dimensional deep algorithm, called hetero-phase volumetric detection (HPVD), that can accept any combination of contrast-phase inputs and with adjustable sensitivity depending on the clinical purpose. We trained HPVD on 771 DCE CT scans to detect HCCs and tested on external 164 positives and 206 controls, respectively. We compare performance against six clinical readers, including two radiologists, two hepato-pancreatico-biliary (HPB) surgeons, and two hepatologists. The area under curve (AUC) of the localization receiver operating characteristic (LROC) for NC-only, NC plus VP, and full DCE CT yielded 0.71, 0.81, 0.89 respectively. At a high sensitivity operating point of 80% on DCE CT, HPVD achieved 97% specificity, which is comparable to measured physician performance. We also demonstrate performance improvements over more typical and less flexible non hetero-phase detectors. Thus, we demonstrate that a single deep learning algorithm can be effectively applied to diverse HCC detection clinical scenarios.

Youbao Tang, Ke Yan, Jinzheng Cai et al.

Measuring lesion size is an important step to assess tumor growth and monitor disease progression and therapy response in oncology image analysis. Although it is tedious and highly time-consuming, radiologists have to work on this task by using RECIST criteria (Response Evaluation Criteria In Solid Tumors) routinely and manually. Even though lesion segmentation may be the more accurate and clinically more valuable means, physicians can not manually segment lesions as now since much more heavy laboring will be required. In this paper, we present a prior-guided dual-path network (PDNet) to segment common types of lesions throughout the whole body and predict their RECIST diameters accurately and automatically. Similar to [1], a click guidance from radiologists is the only requirement. There are two key characteristics in PDNet: 1) Learning lesion-specific attention matrices in parallel from the click prior information by the proposed prior encoder, named click-driven attention; 2) Aggregating the extracted multi-scale features comprehensively by introducing top-down and bottom-up connections in the proposed decoder, named dual-path connection. Experiments show the superiority of our proposed PDNet in lesion segmentation and RECIST diameter prediction using the DeepLesion dataset and an external test set. PDNet learns comprehensive and representative deep image features for our tasks and produces more accurate results on both lesion segmentation and RECIST diameter prediction.

Youbao Tang, Jinzheng Cai, Ke Yan et al.

Accurately segmenting a variety of clinically significant lesions from whole body computed tomography (CT) scans is a critical task on precision oncology imaging, denoted as universal lesion segmentation (ULS). Manual annotation is the current clinical practice, being highly time-consuming and inconsistent on tumor's longitudinal assessment. Effectively training an automatic segmentation model is desirable but relies heavily on a large number of pixel-wise labelled data. Existing weakly-supervised segmentation approaches often struggle with regions nearby the lesion boundaries. In this paper, we present a novel weakly-supervised universal lesion segmentation method by building an attention enhanced model based on the High-Resolution Network (HRNet), named AHRNet, and propose a regional level set (RLS) loss for optimizing lesion boundary delineation. AHRNet provides advanced high-resolution deep image features by involving a decoder, dual-attention and scale attention mechanisms, which are crucial to performing accurate lesion segmentation. RLS can optimize the model reliably and effectively in a weakly-supervised fashion, forcing the segmentation close to lesion boundary. Extensive experimental results demonstrate that our method achieves the best performance on the publicly large-scale DeepLesion dataset and a hold-out test set.

Sicheng Zhao, Quanwei Huang, Youbao Tang et al.

Emotions are usually evoked in humans by images. Recently, extensive research efforts have been dedicated to understanding the emotions of images. In this chapter, we aim to introduce image emotion analysis (IEA) from a computational perspective with the focus on summarizing recent advances and suggesting future directions. We begin with commonly used emotion representation models from psychology. We then define the key computational problems that the researchers have been trying to solve and provide supervised frameworks that are generally used for different IEA tasks. After the introduction of major challenges in IEA, we present some representative methods on emotion feature extraction, supervised classifier learning, and domain adaptation. Furthermore, we introduce available datasets for evaluation and summarize some main results. Finally, we discuss some open questions and future directions that researchers can pursue.

Jieneng Chen, Ke Yan, Yu-Dong Zhang et al.

The boundary of tumors (hepatocellular carcinoma, or HCC) contains rich semantics: capsular invasion, visibility, smoothness, folding and protuberance, etc. Capsular invasion on tumor boundary has proven to be clinically correlated with the prognostic indicator, microvascular invasion (MVI). Investigating tumor boundary semantics has tremendous clinical values. In this paper, we propose the first and novel computational framework that disentangles the task into two components: spatial vertex localization and sequential semantic classification. (1) A HCC tumor segmentor is built for tumor mask boundary extraction, followed by polar transform representing the boundary with radius and angle. Vertex generator is used to produce fixed-length boundary vertices where vertex features are sampled on the corresponding spatial locations. (2) The sampled deep vertex features with positional embedding are mapped into a sequential space and decoded by a multilayer perceptron (MLP) for semantic classification. Extensive experiments on tumor capsule semantics demonstrate the effectiveness of our framework. Mining the correlation between the boundary semantics and MVI status proves the feasibility to integrate this boundary semantics as a valid HCC prognostic biomarker.

Jinzheng Cai, Youbao Tang, Ke Yan et al.

Monitoring treatment response in longitudinal studies plays an important role in clinical practice. Accurately identifying lesions across serial imaging follow-up is the core to the monitoring procedure. Typically this incorporates both image and anatomical considerations. However, matching lesions manually is labor-intensive and time-consuming. In this work, we present deep lesion tracker (DLT), a deep learning approach that uses both appearance- and anatomical-based signals. To incorporate anatomical constraints, we propose an anatomical signal encoder, which prevents lesions being matched with visually similar but spurious regions. In addition, we present a new formulation for Siamese networks that avoids the heavy computational loads of 3D cross-correlation. To present our network with greater varieties of images, we also propose a self-supervised learning (SSL) strategy to train trackers with unpaired images, overcoming barriers to data collection. To train and evaluate our tracker, we introduce and release the first lesion tracking benchmark, consisting of 3891 lesion pairs from the public DeepLesion database. The proposed method, DLT, locates lesion centers with a mean error distance of 7 mm. This is 5% better than a leading registration algorithm while running 14 times faster on whole CT volumes. We demonstrate even greater improvements over detector or similarity-learning alternatives. DLT also generalizes well on an external clinical test set of 100 longitudinal studies, achieving 88% accuracy. Finally, we plug DLT into an automatic tumor monitoring workflow where it leads to an accuracy of 85% in assessing lesion treatment responses, which is only 0.46% lower than the accuracy of manual inputs.

Ke Yan, Jinzheng Cai, Dakai Jin et al.

Radiological images such as computed tomography (CT) and X-rays render anatomy with intrinsic structures. Being able to reliably locate the same anatomical structure across varying images is a fundamental task in medical image analysis. In principle it is possible to use landmark detection or semantic segmentation for this task, but to work well these require large numbers of labeled data for each anatomical structure and sub-structure of interest. A more universal approach would learn the intrinsic structure from unlabeled images. We introduce such an approach, called Self-supervised Anatomical eMbedding (SAM). SAM generates semantic embeddings for each image pixel that describes its anatomical location or body part. To produce such embeddings, we propose a pixel-level contrastive learning framework. A coarse-to-fine strategy ensures both global and local anatomical information are encoded. Negative sample selection strategies are designed to enhance the embedding's discriminability. Using SAM, one can label any point of interest on a template image and then locate the same body part in other images by simple nearest neighbor searching. We demonstrate the effectiveness of SAM in multiple tasks with 2D and 3D image modalities. On a chest CT dataset with 19 landmarks, SAM outperforms widely-used registration algorithms while only taking 0.23 seconds for inference. On two X-ray datasets, SAM, with only one labeled template image, surpasses supervised methods trained on 50 labeled images. We also apply SAM on whole-body follow-up lesion matching in CT and obtain an accuracy of 91%. SAM can also be applied for improving image registration and initializing CNN weights.

Youbao Tang, Ke Yan, Jing Xiao et al.

In clinical trials, one of the radiologists' routine work is to measure tumor sizes on medical images using the RECIST criteria (Response Evaluation Criteria In Solid Tumors). However, manual measurement is tedious and subject to inter-observer variability. We propose a unified framework named SEENet for semi-automatic lesion \textit{SE}gmentation and RECIST \textit{E}stimation on a variety of lesions over the entire human body. The user is only required to provide simple guidance by clicking once near the lesion. SEENet consists of two main parts. The first one extracts the lesion of interest with the one-click guidance, roughly segments the lesion, and estimates its RECIST measurement. Based on the results of the first network, the second one refines the lesion segmentation and RECIST estimation. SEENet achieves state-of-the-art performance in lesion segmentation and RECIST estimation on the large-scale public DeepLesion dataset. It offers a practical tool for radiologists to generate reliable lesion measurements (i.e. segmentation mask and RECIST) with minimal human effort and greatly reduced time.

Youbao Tang, Yuxing Tang, Yingying Zhu et al.

Developing an effective liver and liver tumor segmentation model from CT scans is very important for the success of liver cancer diagnosis, surgical planning and cancer treatment. In this work, we propose a two-stage framework for 2D liver and tumor segmentation. The first stage is a coarse liver segmentation network, while the second stage is an edge enhanced network (E$^2$Net) for more accurate liver and tumor segmentation. E$^2$Net explicitly models complementary objects (liver and tumor) and their edge information within the network to preserve the organ and lesion boundaries. We introduce an edge prediction module in E$^2$Net and design an edge distance map between liver and tumor boundaries, which is used as an extra supervision signal to train the edge enhanced network. We also propose a deep cross feature fusion module to refine multi-scale features from both objects and their edges. E$^2$Net is more easily and efficiently trained with a small labeled dataset, and it can be trained/tested on the original 2D CT slices (resolve resampling error issue in 3D models). The proposed framework has shown superior performance on both liver and liver tumor segmentation compared to several state-of-the-art 2D, 3D and 2D/3D hybrid frameworks.

Yingying Zhu, Youbao Tang, Yuxing Tang et al.

Current deep learning based segmentation models often generalize poorly between domains due to insufficient training data. In real-world clinical applications, cross-domain image analysis tools are in high demand since medical images from different domains are often needed to achieve a precise diagnosis. An important example in radiology is generalizing from non-contrast CT to contrast enhanced CTs. Contrast enhanced CT scans at different phases are used to enhance certain pathologies or organs. Many existing cross-domain image-to-image translation models have been shown to improve cross-domain segmentation of large organs. However, such models lack the ability to preserve fine structures during the translation process, which is significant for many clinical applications, such as segmenting small calcified plaques in the aorta and pelvic arteries. In order to preserve fine structures during medical image translation, we propose a patch-based model using shared latent variables from a Gaussian mixture model. We compare our image translation framework to several state-of-the-art methods on cross-domain image translation and show our model does a better job preserving fine structures. The superior performance of our model is verified by performing two tasks with the translated images - detection and segmentation of aortic plaques and pancreas segmentation. We expect the utility of our framework will extend to other problems beyond segmentation due to the improved quality of the generated images and enhanced ability to preserve small structures.

Jia Liang, Yuxing Tang, Youbao Tang et al.

Dual-energy (DE) chest radiography provides the capability of selectively imaging two clinically relevant materials, namely soft tissues, and osseous structures, to better characterize a wide variety of thoracic pathology and potentially improve diagnosis in posteroanterior (PA) chest radiographs. However, DE imaging requires specialized hardware and a higher radiation dose than conventional radiography, and motion artifacts sometimes happen due to involuntary patient motion. In this work, we learn the mapping between conventional radiographs and bone suppressed radiographs. Specifically, we propose to utilize two variations of generative adversarial networks (GANs) for image-to-image translation between conventional and bone suppressed radiographs obtained by DE imaging technique. We compare the effectiveness of training with patient-wisely paired and unpaired radiographs. Experiments show both training strategies yield "radio-realistic'' radiographs with suppressed bony structures and few motion artifacts on a hold-out test set. While training with paired images yields slightly better performance than that of unpaired images when measuring with two objective image quality metrics, namely Structural Similarity Index (SSIM) and Peak Signal-to-Noise Ratio (PSNR), training with unpaired images demonstrates better generalization ability on unseen anteroposterior (AP) radiographs than paired training.

Vatsal Agarwal, Youbao Tang, Jing Xiao et al.

Lesion segmentation in medical imaging serves as an effective tool for assessing tumor sizes and monitoring changes in growth. However, not only is manual lesion segmentation time-consuming, but it is also expensive and requires expert radiologist knowledge. Therefore many hospitals rely on a loose substitute called response evaluation criteria in solid tumors (RECIST). Although these annotations are far from precise, they are widely used throughout hospitals and are found in their picture archiving and communication systems (PACS). Therefore, these annotations have the potential to serve as a robust yet challenging means of weak supervision for training full lesion segmentation models. In this work, we propose a weakly-supervised co-segmentation model that first generates pseudo-masks from the RECIST slices and uses these as training labels for an attention-based convolutional neural network capable of segmenting common lesions from a pair of CT scans. To validate and test the model, we utilize the DeepLesion dataset, an extensive CT-scan lesion dataset that contains 32,735 PACS bookmarked images. Extensive experimental results demonstrate the efficacy of our co-segmentation approach for lesion segmentation with a mean Dice coefficient of 90.3%.

Vatsal Agarwal, Youbao Tang, Jing Xiao et al.

Lesion segmentation on computed tomography (CT) scans is an important step for precisely monitoring changes in lesion/tumor growth. This task, however, is very challenging since manual segmentation is prohibitively time-consuming, expensive, and requires professional knowledge. Current practices rely on an imprecise substitute called response evaluation criteria in solid tumors (RECIST). Although these markers lack detailed information about the lesion regions, they are commonly found in hospitals' picture archiving and communication systems (PACS). Thus, these markers have the potential to serve as a powerful source of weak-supervision for 2D lesion segmentation. To approach this problem, this paper proposes a convolutional neural network (CNN) based weakly-supervised lesion segmentation method, which first generates the initial lesion masks from the RECIST measurements and then utilizes co-segmentation to leverage lesion similarities and refine the initial masks. In this work, an attention-based co-segmentation model is adopted due to its ability to learn more discriminative features from a pair of images. Experimental results on the NIH DeepLesion dataset demonstrate that the proposed co-segmentation approach significantly improves lesion segmentation performance, e.g the Dice score increases about 4.0% (from 85.8% to 89.8%).

Yuxing Tang, Youbao Tang, Veit Sandfort et al.

In this work, we exploit the unsupervised domain adaptation problem for radiology image interpretation across domains. Specifically, we study how to adapt the disease recognition model from a labeled source domain to an unlabeled target domain, so as to reduce the effort of labeling each new dataset. To address the shortcoming of cross-domain, unpaired image-to-image translation methods which typically ignore class-specific semantics, we propose a task-driven, discriminatively trained, cycle-consistent generative adversarial network, termed TUNA-Net. It is able to preserve 1) low-level details, 2) high-level semantic information and 3) mid-level feature representation during the image-to-image translation process, to favor the target disease recognition task. The TUNA-Net framework is general and can be readily adapted to other learning tasks. We evaluate the proposed framework on two public chest X-ray datasets for pneumonia recognition. The TUNA-Net model can adapt labeled adult chest X-rays in the source domain such that they appear as if they were drawn from pediatric X-rays in the unlabeled target domain, while preserving the disease semantics. Extensive experiments show the superiority of the proposed method as compared to state-of-the-art unsupervised domain adaptation approaches. Notably, TUNA-Net achieves an AUC of 96.3% for pediatric pneumonia classification, which is very close to that of the supervised approach (98.1%), but without the need for labels on the target domain.

Ke Yan, Youbao Tang, Yifan Peng et al.

When reading medical images such as a computed tomography (CT) scan, radiologists generally search across the image to find lesions, characterize and measure them, and then describe them in the radiological report. To automate this process, we propose a multitask universal lesion analysis network (MULAN) for joint detection, tagging, and segmentation of lesions in a variety of body parts, which greatly extends existing work of single-task lesion analysis on specific body parts. MULAN is based on an improved Mask R-CNN framework with three head branches and a 3D feature fusion strategy. It achieves the state-of-the-art accuracy in the detection and tagging tasks on the DeepLesion dataset, which contains 32K lesions in the whole body. We also analyze the relationship between the three tasks and show that tag predictions can improve detection accuracy via a score refinement layer.

Yuxing Tang, Youbao Tang, Mei Han et al.

Being one of the most common diagnostic imaging tests, chest radiography requires timely reporting of potential findings in the images. In this paper, we propose an end-to-end architecture for abnormal chest X-ray identification using generative adversarial one-class learning. Unlike previous approaches, our method takes only normal chest X-ray images as input. The architecture is composed of three deep neural networks, each of which learned by competing while collaborating among them to model the underlying content structure of the normal chest X-rays. Given a chest X-ray image in the testing phase, if it is normal, the learned architecture can well model and reconstruct the content; if it is abnormal, since the content is unseen in the training phase, the model would perform poorly in its reconstruction. It thus enables distinguishing abnormal chest X-rays from normal ones. Quantitative and qualitative experiments demonstrate the effectiveness and efficiency of our approach, where an AUC of 0.841 is achieved on the challenging NIH Chest X-ray dataset in a one-class learning setting, with the potential in reducing the workload for radiologists.

Youbao Tang, Ke Yan, Yuxing Tang et al.

Automatic lesion detection from computed tomography (CT) scans is an important task in medical imaging analysis. It is still very challenging due to similar appearances (e.g. intensity and texture) between lesions and other tissues, making it especially difficult to develop a universal lesion detector. Instead of developing a specific-type lesion detector, this work builds a Universal Lesion Detector (ULDor) based on Mask R-CNN, which is able to detect all different kinds of lesions from whole body parts. As a state-of-the-art object detector, Mask R-CNN adds a branch for predicting segmentation masks on each Region of Interest (RoI) to improve the detection performance. However, it is almost impossible to manually annotate a large-scale dataset with pixel-level lesion masks to train the Mask R-CNN for lesion detection. To address this problem, this work constructs a pseudo mask for each lesion region that can be considered as a surrogate of the real mask, based on which the Mask R-CNN is employed for lesion detection. On the other hand, this work proposes a hard negative example mining strategy to reduce the false positives for improving the detection performance. Experimental results on the NIH DeepLesion dataset demonstrate that the ULDor is enhanced using pseudo masks and the proposed hard negative example mining strategy and achieves a sensitivity of 86.21% with five false positives per image.

Youbao Tang, Jinzheng Cai, Le Lu et al.

Automated lesion segmentation from computed tomography (CT) is an important and challenging task in medical image analysis. While many advancements have been made, there is room for continued improvements. One hurdle is that CT images can exhibit high noise and low contrast, particularly in lower dosages. To address this, we focus on a preprocessing method for CT images that uses stacked generative adversarial networks (SGAN) approach. The first GAN reduces the noise in the CT image and the second GAN generates a higher resolution image with enhanced boundaries and high contrast. To make up for the absence of high quality CT images, we detail how to synthesize a large number of low- and high-quality natural images and use transfer learning with progressively larger amounts of CT images. We apply both the classic GrabCut method and the modern holistically nested network (HNN) to lesion segmentation, testing whether SGAN can yield improved lesion segmentation. Experimental results on the DeepLesion dataset demonstrate that the SGAN enhancements alone can push GrabCut performance over HNN trained on original images. We also demonstrate that HNN + SGAN performs best compared against four other enhancement methods, including when using only a single GAN.

Jinzheng Cai, Youbao Tang, Le Lu et al.

Volumetric lesion segmentation from computed tomography (CT) images is a powerful means to precisely assess multiple time-point lesion/tumor changes. However, because manual 3D segmentation is prohibitively time consuming, current practices rely on an imprecise surrogate called response evaluation criteria in solid tumors (RECIST). Despite their coarseness, RECIST markers are commonly found in current hospital picture and archiving systems (PACS), meaning they can provide a potentially powerful, yet extraordinarily challenging, source of weak supervision for full 3D segmentation. Toward this end, we introduce a convolutional neural network (CNN) based weakly supervised slice-propagated segmentation (WSSS) method to 1) generate the initial lesion segmentation on the axial RECIST-slice; 2) learn the data distribution on RECIST-slices; 3) extrapolate to segment the whole lesion slice by slice to finally obtain a volumetric segmentation. To validate the proposed method, we first test its performance on a fully annotated lymph node dataset, where WSSS performs comparably to its fully supervised counterparts. We then test on a comprehensive lesion dataset with 32,735 RECIST marks, where we report a mean Dice score of 92% on RECIST-marked slices and 76% on the entire 3D volumes.

Youbao Tang, Adam P. Harrison, Mohammadhadi Bagheri et al.

Response evaluation criteria in solid tumors (RECIST) is the standard measurement for tumor extent to evaluate treatment responses in cancer patients. As such, RECIST annotations must be accurate. However, RECIST annotations manually labeled by radiologists require professional knowledge and are time-consuming, subjective, and prone to inconsistency among different observers. To alleviate these problems, we propose a cascaded convolutional neural network based method to semi-automatically label RECIST annotations and drastically reduce annotation time. The proposed method consists of two stages: lesion region normalization and RECIST estimation. We employ the spatial transformer network (STN) for lesion region normalization, where a localization network is designed to predict the lesion region and the transformation parameters with a multi-task learning strategy. For RECIST estimation, we adapt the stacked hourglass network (SHN), introducing a relationship constraint loss to improve the estimation precision. STN and SHN can both be learned in an end-to-end fashion. We train our system on the DeepLesion dataset, obtaining a consensus model trained on RECIST annotations performed by multiple radiologists over a multi-year period. Importantly, when judged against the inter-reader variability of two additional radiologist raters, our system performs more stably and with less variability, suggesting that RECIST annotations can be reliably obtained with reduced labor and time.

Dakai Jin, Ziyue Xu, Youbao Tang et al.

Data availability plays a critical role for the performance of deep learning systems. This challenge is especially acute within the medical image domain, particularly when pathologies are involved, due to two factors: 1) limited number of cases, and 2) large variations in location, scale, and appearance. In this work, we investigate whether augmenting a dataset with artificially generated lung nodules can improve the robustness of the progressive holistically nested network (P-HNN) model for pathological lung segmentation of CT scans. To achieve this goal, we develop a 3D generative adversarial network (GAN) that effectively learns lung nodule property distributions in 3D space. In order to embed the nodules within their background context, we condition the GAN based on a volume of interest whose central part containing the nodule has been erased. To further improve realism and blending with the background, we propose a novel multi-mask reconstruction loss. We train our method on over 1000 nodules from the LIDC dataset. Qualitative results demonstrate the effectiveness of our method compared to the state-of-art. We then use our GAN to generate simulated training images where nodules lie on the lung border, which are cases where the published P-HNN model struggles. Qualitative and quantitative results demonstrate that armed with these simulated images, the P-HNN model learns to better segment lung regions under these challenging situations. As a result, our system provides a promising means to help overcome the data paucity that commonly afflicts medical imaging.

Jinzheng Cai, Youbao Tang, Le Lu et al.

Volumetric lesion segmentation via medical imaging is a powerful means to precisely assess multiple time-point lesion/tumor changes. Because manual 3D segmentation is prohibitively time consuming and requires radiological experience, current practices rely on an imprecise surrogate called response evaluation criteria in solid tumors (RECIST). Despite their coarseness, RECIST marks are commonly found in current hospital picture and archiving systems (PACS), meaning they can provide a potentially powerful, yet extraordinarily challenging, source of weak supervision for full 3D segmentation. Toward this end, we introduce a convolutional neural network based weakly supervised self-paced segmentation (WSSS) method to 1) generate the initial lesion segmentation on the axial RECIST-slice; 2) learn the data distribution on RECIST-slices; 3) adapt to segment the whole volume slice by slice to finally obtain a volumetric segmentation. In addition, we explore how super-resolution images (2~5 times beyond the physical CT imaging), generated from a proposed stacked generative adversarial network, can aid the WSSS performance. We employ the DeepLesion dataset, a comprehensive CT-image lesion dataset of 32,735 PACS-bookmarked findings, which include lesions, tumors, and lymph nodes of varying sizes, categories, body regions and surrounding contexts. These are drawn from 10,594 studies of 4,459 patients. We also validate on a lymph-node dataset, where 3D ground truth masks are available for all images. For the DeepLesion dataset, we report mean Dice coefficients of 93% on RECIST-slices and 76% in 3D lesion volumes. We further validate using a subjective user study, where an experienced radiologist accepted our WSSS-generated lesion segmentation results with a high probability of 92.4%.

Youbao Tang, Xiangqian Wu

This paper proposes a novel saliency detection method by combining region-level saliency estimation and pixel-level saliency prediction with CNNs (denoted as CRPSD). For pixel-level saliency prediction, a fully convolutional neural network (called pixel-level CNN) is constructed by modifying the VGGNet architecture to perform multi-scale feature learning, based on which an image-to-image prediction is conducted to accomplish the pixel-level saliency detection. For region-level saliency estimation, an adaptive superpixel based region generation technique is first designed to partition an image into regions, based on which the region-level saliency is estimated by using a CNN model (called region-level CNN). The pixel-level and region-level saliencies are fused to form the final salient map by using another CNN (called fusion CNN). And the pixel-level CNN and fusion CNN are jointly learned. Extensive quantitative and qualitative experiments on four public benchmark datasets demonstrate that the proposed method greatly outperforms the state-of-the-art saliency detection approaches.

Youbao Tang, Xiangqian Wu, Wei Bu

This paper proposes a novel saliency detection method by developing a deeply-supervised recurrent convolutional neural network (DSRCNN), which performs a full image-to-image saliency prediction. For saliency detection, the local, global, and contextual information of salient objects is important to obtain a high quality salient map. To achieve this goal, the DSRCNN is designed based on VGGNet-16. Firstly, the recurrent connections are incorporated into each convolutional layer, which can make the model more powerful for learning the contextual information. Secondly, side-output layers are added to conduct the deeply-supervised operation, which can make the model learn more discriminative and robust features by effecting the intermediate layers. Finally, all of the side-outputs are fused to integrate the local and global information to get the final saliency detection results. Therefore, the DSRCNN combines the advantages of recurrent convolutional neural networks and deeply-supervised nets. The DSRCNN model is tested on five benchmark datasets, and experimental results demonstrate that the proposed method significantly outperforms the state-of-the-art saliency detection approaches on all test datasets.