Jiaxiang Ren


jiaxren AT cs.stonybrook.edu


Curriculum Vitae | GitHub


I am a CS PhD student at the Department of Computer Science of Stony Brook University, under the supervision of Prof. Haibin Ling and Prof. Yingtian Pan. My research interests mainly focus on Computer Vision and pattern recognition.
Previously, I was an algorithm engineer in Ping An Technology, working on computer vision, machine learning and deep learning. I obtained the BEng and MS degree in Computer Science and Technology from Tongji University, advised by Prof. Shengjie Zhao.



Self-Supervised Bulk Motion Artifact Removal in Optical Coherence Tomography Angiography


Jiaxiang Ren, Kicheon Park, Yingtian Pan, and Haibin Ling


IEEE / CVF Computer Vision and Pattern Recognition Conference(CVPR), 2022


Paper | Code | Abstract

Optical coherence tomography angiography (OCTA) is an important imaging modality in many bioengineering tasks. The image quality of OCTA, however, is often hurt by Bulk Motion Artifacts (BMA), which are due to micromotion of subjects and typically appear as bright stripes surrounded by blur areas. State-of-the-art BMA handling solutions usually treat the problem as an image inpainting one with deep neural network algorithms. These solutions, however, require numerous training samples with nontrivial annotation. Nevertheless, this context-based inpainting model has limited correction capability because it discards the rich structural and appearance information carried in the BMA stripe region. To address these issues, in this paper we propose a self-supervised content-aware BMA recover model. First, the gradient-based structural information and appearance feature are extracted from the BMA area and injected into the model to capture more connectivity. Second, with easily collected defective masks, the model is trained in a self-supervised manner that only the clear areas are for training while the BMA areas for inference. With structural information and appearance feature from noisy image as references, our model could correct larger BMA and produce better visualizing result. Only 2D images with defective masks are involved so our method is more efficient. Experiments on OCTA of mouse cortex demonstrate that our model could correct most BMA with extremely large sizes and inconsistent intensities while existing methods fail.


Dynamic 3D imaging of cerebral blood flow in awake mice using self supervised-learning-enhanced optical coherence Doppler tomography


Yingtian Pan, Kicheon Park, Jiaxiang Ren, Nora D. Volkow, Haibin Ling, Alan P. Koretsky, Congwu Du


Nature Communications Biology, 2023 (IF: 6.548)


Code | Abstract

Cerebral blood flow (CBF) is widely used to assess brain function in normal and diseased states. However, most preclinical CBF imaging is performed under anesthesia, thus brain functional changes are confounded by anesthetic effects. Due to motion artifacts and high background noise, high spatiotemporal-resolution CBF imaging of awake animals is challenging, particularly for Doppler-based flow imaging modalities. Here, we report ultrahigh-resolution optical coherence Doppler tomography (μODT) for 3D imaging of CBF velocity (CBFv) network dynamics in head- restrained awake-behaving animals by developing self-supervised deep-learning approaches for effective image denoising and motion-artifact removal.


Osteoporosis Prescreening and Bone Mineral Density Prediction using Dental Panoramic Radiographs


Yasha Singh, Vivek Atulkar, Jiaxiang Ren, Jie Yang, Heng Fan, Longin Jan Latecki, Haibin Ling


43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2021


Paper | Abstract

Recent studies have shown that Dental Panoramic Radiograph (DPR) images have great potential for prescreening of osteoporosis given the high degree of correlation between the bone density and trabecular bone structure. Most of the research works in these area had used pretrained models for feature extraction and classification with good success. However, when the size of the data set is limited it becomes difficult to use these pretrained networks and gain high confidence scores. In this paper, we evaluated the diagnostic performance of deep convolutional neural networks (DCNN)- based computer-assisted diagnosis (CAD) system in the detection of osteoporosis on panoramic radiographs, through a comparison with diagnoses made by oral and maxillofacial radiologists. With the available labelled dataset of 70 images, results were reproduced for the preliminary study model. Furthermore, the model performance was enhanced using different computer vision techniques. Specifically, the age meta data available for each patient was leveraged to obtain more accurate predictions. Lastly, we tried to leverage these images, ages and osteoporotic labels to create a neural network based regression model and predict the Bone Mineral Density (BMD) value for each patient. Experimental results showed that the proposed CAD system was in high accord with experienced oral and maxillofacial radiologists in detecting osteoporosis and achieved 87.86% accuracy.


Detection of Trabecular Landmarks for Osteoporosis Prescreening in Dental Panoramic Radiographs


Jiaxiang Ren, Heng Fan, Jie Yang, and Haibin Ling


42rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2020


Paper | Abstract

Dental panoramic radiography (DPR) images have recently attracted increasing attention in osteoporosis analysis because of their inner correlation. Many approaches leverage machine learning techniques (e.g. , deep convolutional neural networks (CNNs)) to study DPR images of a patient to provide initial analysis of osteoporosis, which demonstrates promising results and significantly reduces financial cost. However, these methods heavily rely on the trabecula landmarks of DPR images that requires a large amount of manual annotations by dentist, and thus are limited in practical application. Addressing this issue, we propose to automatically detect trabecular landmarks in DPR images. In specific, we first apply CNNs-based detector for trabecular landmark detection and analyze its limitations. Using CNNs-based detection as a baseline, we then introduce a statistic shape model (SSM) for trabecular landmark detection by taking advantage of spatial distribution prior of trabecular landmarks in DPR images and their structural relations. In experiment on 108 images, our solution outperforms CNNs- based detector. Moreover, compared to CNN-based detectors, our method avoids the needs of vast training samples, which is more practical in application.


Enhanced Sparse Subspace Clustering


Jiaxiang Ren, Shengjie Zhao, Kai Yang, and Brian Zhao


IEEE International Conference on Multimedia & Expo Workshops(ICMEW), 2017


Project | Paper | Slides | Bibtex | Abstract

High-dimensional data are ubiquitous in most real-world research areas, such as machine learning, image processing and so on. Actually, high-dimensional data that belong to the same classes tend to gather in their own low-dimensional subspaces. Recently, many subspace recovery algorithms based on the sparse representation such as Sparse Subspace Clustering (SSC), Low-Rank Representation (LRR) and their variants are proposed to address the subspace clustering problems.
In this paper, we propose a novel clustering method based on SSC, called Enhanced Sparse Subspace Clustering (ESSC), to deal with complicated face images under variant expressions, illuminations or disguises. Assuming that the variant expressions or disguises of the face images are sharable, we introduce the adaptive difference dictionary to extend the simple linear combination of face images in SSC with the combination of the specific features and the common features. Thus both of the accuracy and generalization of clustering are improved simultaneously.The experimental results on the AR face databases show that the proposed ESSC not only makes up to 9.0% improvements on accuracy compared with SSC, but also is more robust and scalable for dealing with larger face clustering problems (up to 400 samples from 100 subjects) under variant expressions, disguises or illuminations.

  title={A novel and robust face clustering method via adaptive difference dictionary},
  author={Jiaxiang Ren and Shengjie Zhao and Kai Yang and Brian Nlong Zhao},
  booktitle={IEEE International Conference on Multimedia & Expo Workshops (ICMEW)},



Tuberculosis Screening


Jiaxiang Ren, Lingyun Huang





Tuberculosis is a serious disease that affects lots of people in developing countries. To treat TB patients, an accurate diagnosis is needed firstly. X-ray radiography is often used for TB screening among high risk populations but manual examining cheXray is time consuming. For improving efficiency and reducing screening cost, we develop a series models to classify normal and TB patients' cheXray images, find and mark related infections within chest.
According to the distribution of our TB dataset, we experiment with different segmentation and detection models, such as attention-based DANet and anchor-free FCOS. Our final model is specifically designed for small infections, which improve recall with insignificant loss in precision.


RSNA Pneumonia Detection Challenge (Kaggle)


Jiaxiang Ren, Liangxin Gao, Yanbo Zhang


Kaggle Competition (Rank: 39/1499, Top 3%)


Project | Abstract

Pneumonia accounts for over 15% of all deaths of children under 5 years old internationally. In 2015, 920,000 children under the age of 5 died from the disease. To improve the efficiency and reach of diagnostic services, the Radiological Society of North America (RSNA) has reached out to machine learning community of Kaggle and collaborated with the US National Institutes of Health, The Society of Thoracic Radiology, and MD.ai to develop a rich dataset for this challenge. Our final model is mainly based on keras-YOLOv3 with Hard Negative Mining(HNM) to reduce false positive rates. Among 1499 competitors, our team finally achieved the 39th place (Top 3%).


Fine-grained Plants Categorization


Jiaxiang Ren, Aimee Liu, Jin Ma





Atractylodes lancea is a high-demand Chinese herbal medicine (annual output > 1,000 tons). The main medicinal part of Atractylodes lancea is the root, which is very similar to the root of Atractylodes macrocephala. Only experienced personnel could distinguish them correctly. As a result, it is difficult to meet the requirement of production and purity at the same time. Separating these two similar plants during cultivating is a solution to increase the yield of high-purity medicine. However, when cultivating, most plants are without distinguishable flowers as illustrated in the left image. To this end, we take advantage of transfer-learning to solve this fine-grained categorization problem.
After analysing all the samples in our herb dataset, we process image into affordable size while its discriminative features are preserved. Thereafter, we use transfer-learning and other regular tricks to train our model.
*This illustration image is from Medicinal Plant Images Database.


Improvement of 3D Semantic Completion from a Single Depth Image


Jiaxiang Ren, Shengjie Zhao


Graduate Research


Project | Abstract

Project aims at completing the 3D semantic scene from the depth images. We introduced a post-processing module to improve the Average Precision (AP) and Intersection-over-union (IoU). The aggregated end-to-end network improves the AP by 6.1% and IoU by 2.3%.


The Research on the Sparse-based Subspace Clustering Algorithm in High-dimensional Data


Jiaxiang Ren, Shengjie Zhao


Master Thesis


Project | Abstract

High-dimensional data is ubiquitous in machine learning, pattern recognition, image processing, computer vision, and bioinformatics. It causes “Curse of dimensionality”, which not only increases the cost but also introduces noise data. However, high-dimensional data is usually gathered in the low-dimensional subspaces. This property brings hope for us to solve the “dimensional curse” problem. This paper research on the sparse subspace clustering algorithm. The effectiveness of the proposed algorithm for complex noises is verified through experiments.
By constructing an adaptive difference dictionary, the enhanced algorithm can better restore the structure of the underlying subspace and improve the accuracy. The randomized algorithm is optional to the adaptive difference dictionary, which can greatly increase the speed of calculation when degrades little performance. So it is suitable for environments with limited computing power.



web page statistics from GoStats