Hanqing Chao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Hanqing Chao has a total of 46 co-authors. Here are the top co-authors by number of shared publications:
- Pingkun Yan
Rensselaer Polytechnic Institute
9 shared publications - Junping Zhang
Fudan University
6 shared publications - Ge Wang
Rensselaer Polytechnic Institute
4 shared publications - Jiajin Zhang
Department of Biomedical Engineering and the Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, NY, USA
4 shared publications - Mannudeep K. Kalra
Department of Radiology, Massachusetts General Hospital, Boston, MA 02114, USA
4 shared publications - Diego Machado Reyes
Rensselaer Polytechnic Institute
3 shared publications
Latest Publications
- Open Access
- 15 April 2024
- Published by MDPI AG in Journal of Personalized Medicine
- Vol. 14 (4) , 421
Identifying Progression-Specific Alzheimer’s Subtypes Using Multimodal Transformer
Alzheimer’s disease (AD) is the most prevalent neurodegenerative disease, yet its current treatments are limited to stopping disease progression. Moreover, the effectiveness of these treatments remains uncertain due to the heterogeneity of the disease. Therefore, it is essential to identify disease subtypes at a very early stage. Current data-driven approaches can be used to classify subtypes during later stages of AD or related disorders, but making predictions in the asymptomatic or prodromal stage is challenging. Furthermore, the classifications of most existing models lack explainability, and these models rely solely on a single modality for assessment, limiting the scope of their analysis. Thus, we propose a multimodal framework that utilizes early-stage indicators, including imaging, genetics, and clinical assessments, to classify AD patients into progression-specific subtypes at an early stage. In our framework, we introduce a tri-modal co-attention mechanism (Tri-COAT) to explicitly capture cross-modal feature associations. Data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (slow progressing = 177, intermediate = 302, and fast = 15) were used to train and evaluate Tri-COAT using a 10-fold stratified cross-testing approach. Our proposed model outperforms baseline models and sheds light on essential associations across multimodal features supported by known biological mechanisms. The multimodal design behind Tri-COAT allows it to achieve the highest classification area under the receiver operating characteristic curve while simultaneously providing interpretability to the model predictions through the co-attention mechanism.
- 15 April 2024
- Published by MDPI AG in Journal of Personalized Medicine
- Vol. 14 (4) , 421
Open AccessIdentifying Progression-Specific Alzheimer’s Subtypes Using Multimodal Transformer
Alzheimer’s disease (AD) is the most prevalent neurodegenerative disease, yet its current treatments are limited to stopping disease progression. Moreover, the effectiveness of these treatments remains uncertain due to the heterogeneity of the disease. Therefore, it is essential to identify disease subtypes at a very early stage. Current data-driven approaches can be used to classify subtypes during later stages of AD or related disorders, but making predictions in the asymptomatic or prodromal stage is challenging. Furthermore, the classifications of most existing models lack explainability, and these models rely solely on a single modality for assessment, limiting the scope of their analysis. Thus, we propose a multimodal framework that utilizes early-stage indicators, including imaging, genetics, and clinical assessments, to classify AD patients into progression-specific subtypes at an early stage. In our framework, we introduce a tri-modal co-attention mechanism (Tri-COAT) to explicitly capture cross-modal feature associations. Data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (slow progressing = 177, intermediate = 302, and fast = 15) were used to train and evaluate Tri-COAT using a 10-fold stratified cross-testing approach. Our proposed model outperforms baseline models and sheds light on essential associations across multimodal features supported by known biological mechanisms. The multimodal design behind Tri-COAT allows it to achieve the highest classification area under the receiver operating characteristic curve while simultaneously providing interpretability to the model predictions through the co-attention mechanism.
- Open Access
- 1 September 2023
- Published by KEAI Publishing in Meta-Radiology
- Vol. 1 (2)
Multimodal radiology AI
- 1 September 2023
- Published by KEAI Publishing in Meta-Radiology
- Vol. 1 (2)
Open AccessMultimodal radiology AI
- 13 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 32 (10577149) , 1272-1284
Toward Adversarial Robustness in Unlabeled Target Domains
In the past several years, various adversarial training (AT) approaches have been invented to robustify deep learning model against adversarial attacks. However, mainstream AT methods assume the training and testing data are drawn from the same distribution and the training data are annotated. When the two assumptions are violated, existing AT methods fail because either they cannot pass knowledge learnt from a source domain to an unlabeled target domain or they are confused by the adversarial samples in that unlabeled space. In this paper, we first point out this new and challenging problem— adversarial training in unlabeled target domain . We then propose a novel framework named Unsupervised Cross-domain Adversarial Training (UCAT) to address this problem. UCAT effectively leverages the knowledge of the labeled source domain to prevent the adversarial samples from misleading the training process, under the guidance of automatically selected high quality pseudo labels of the unannotated target domain data together with the discriminative and robust anchor representations of the source domain data. The experiments on four public benchmarks show that models trained with UCAT can achieve both high accuracy and strong robustness. The effectiveness of the proposed components is demonstrated through a large set of ablation studies. The source code is publicly available at https://github.com/DIAL-RPI/UCAT.
- 13 February 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 32 (10577149) , 1272-1284
Toward Adversarial Robustness in Unlabeled Target Domains
In the past several years, various adversarial training (AT) approaches have been invented to robustify deep learning model against adversarial attacks. However, mainstream AT methods assume the training and testing data are drawn from the same distribution and the training data are annotated. When the two assumptions are violated, existing AT methods fail because either they cannot pass knowledge learnt from a source domain to an unlabeled target domain or they are confused by the adversarial samples in that unlabeled space. In this paper, we first point out this new and challenging problem— adversarial training in unlabeled target domain . We then propose a novel framework named Unsupervised Cross-domain Adversarial Training (UCAT) to address this problem. UCAT effectively leverages the knowledge of the labeled source domain to prevent the adversarial samples from misleading the training process, under the guidance of automatically selected high quality pseudo labels of the unannotated target domain data together with the discriminative and robust anchor representations of the source domain data. The experiments on four public benchmarks show that models trained with UCAT can achieve both high accuracy and strong robustness. The effectiveness of the proposed components is demonstrated through a large set of ablation studies. The source code is publicly available at https://github.com/DIAL-RPI/UCAT.
- 27 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE EMBS International Conference on Biomedical & Health Informatics (BHI)
- Vol. 2022, 01-04
Genomics transformer for diagnosing Parkinson's disease
Parkinson's disease (PD) is the second most common neurodegenerative disease and presents a complex etiology with genomic and environmental factors and no recognized cures. Genotype data, such as single nucleotide polymorphisms (SNPs), could be used as a prodromal factor for early detection of PD. However, the polygenic nature of PD presents a challenge as the complex relationships between SNPs towards disease development are difficult to model. Traditional assessment methods such as polygenic risk scores and machine learning approaches struggle to capture the complex interactions present in the genotype data, thus limiting their discriminative capabilities in diagnosis. On the other hand, DL models are better suited for this task. Nevertheless, they encounter difficulties of their own such as a lack of interpretability. To overcome these limitations, in this work, a novel transformer encoder-based model is introduced to classify PD patients from healthy controls based on their genotype. This method is designed to effectively model complex global feature interactions and enable increased interpretability through the learned attention scores. The proposed framework outperformed traditional machine learning and deep learning baseline models. Moreover, visualization of the learned SNP-SNP associations provides not only interpretability to the model but also valuable insights into the biochemical pathways underlying PD development, which are corroborated by pathway enrichment analysis. Our results suggest novel SNP interactions to be further studied in wet lab and clinical settings.
- 27 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE EMBS International Conference on Biomedical & Health Informatics (BHI)
- Vol. 2022, 01-04
Genomics transformer for diagnosing Parkinson's disease
Parkinson's disease (PD) is the second most common neurodegenerative disease and presents a complex etiology with genomic and environmental factors and no recognized cures. Genotype data, such as single nucleotide polymorphisms (SNPs), could be used as a prodromal factor for early detection of PD. However, the polygenic nature of PD presents a challenge as the complex relationships between SNPs towards disease development are difficult to model. Traditional assessment methods such as polygenic risk scores and machine learning approaches struggle to capture the complex interactions present in the genotype data, thus limiting their discriminative capabilities in diagnosis. On the other hand, DL models are better suited for this task. Nevertheless, they encounter difficulties of their own such as a lack of interpretability. To overcome these limitations, in this work, a novel transformer encoder-based model is introduced to classify PD patients from healthy controls based on their genotype. This method is designed to effectively model complex global feature interactions and enable increased interpretability through the learned attention scores. The proposed framework outperformed traditional machine learning and deep learning baseline models. Moreover, visualization of the learned SNP-SNP associations provides not only interpretability to the model but also valuable insights into the biochemical pathways underlying PD development, which are corroborated by pathway enrichment analysis. Our results suggest novel SNP interactions to be further studied in wet lab and clinical settings.
- 16 September 2022
- Published by Springer Nature
- p. 427-436
Regression Metric Loss: Learning a Semantic Representation Space for Medical Images
- 16 September 2022
- Published by Springer Nature
- p. 427-436
Regression Metric Loss: Learning a Semantic Representation Space for Medical Images
- Open Access
- 29 July 2022
- Published by Springer Nature in Research Square
Quantifying Trustworthiness of Explainability in Medical AI
Saliency visualization methods help explain artificial intelligence (AI) models and build the trust of AI-driven medical image analysis applications. However, the trustworthiness of the generated explanations is often overlooked. Our article demonstrates that the vulnerabilities of such explanations to subtle perturbations of the input can lead to misleading results. More importantly, we show that vulnerabilities can be exploited without knowing the details of an AI model. We then propose criteria and methods to evaluate the trustworthiness of saliency maps and report a series of systematic evaluation and reader studies of widely adopted deep neural networks on large scale public datasets. The results show that the saliency may not be relevant to the model output, and it can be tampered by maintaining the model output without knowing the model specifics. Evidence suggests that AI researchers, practitioners, and authoritative agencies in the medical domain should use caution when explaining AI models.
- 29 July 2022
- Published by Springer Nature in Research Square
Open AccessQuantifying Trustworthiness of Explainability in Medical AI
Saliency visualization methods help explain artificial intelligence (AI) models and build the trust of AI-driven medical image analysis applications. However, the trustworthiness of the generated explanations is often overlooked. Our article demonstrates that the vulnerabilities of such explanations to subtle perturbations of the input can lead to misleading results. More importantly, we show that vulnerabilities can be exploited without knowing the details of an AI model. We then propose criteria and methods to evaluate the trustworthiness of saliency maps and report a series of systematic evaluation and reader studies of widely adopted deep neural networks on large scale public datasets. The results show that the saliency may not be relevant to the model output, and it can be tampered by maintaining the model output without knowing the model specifics. Evidence suggests that AI researchers, practitioners, and authoritative agencies in the medical domain should use caution when explaining AI models.
- Open Access
- 23 April 2022
- Published by SPIE-Intl Soc Optical Eng in Neurophotonics
- Vol. 9 (04) , 041406-15
Deep learning-based motion artifact removal in functional near-infrared spectroscopy
Functional near-infrared spectroscopy (fNIRS) is a noninvasive neuroimaging technique to monitor brain activity indirectly. It measures the intensity of near-infrared light diffusely scattered through cortical tissues to quantify changes in oxyhemoglobin concentration (HbO) and deoxyhemoglobin concentration (HbR) with respect to cortical regions.1 – 4 HbO and HbR time series can reflect changes in neurovascular coupling and hence, neuronal activity. fNIRS has been widely used in cognitive,5 – 9 motor skill studies,10 – 15 and brain–computer interface technique.16
- 23 April 2022
- Published by SPIE-Intl Soc Optical Eng in Neurophotonics
- Vol. 9 (04) , 041406-15
Open AccessDeep learning-based motion artifact removal in functional near-infrared spectroscopy
Functional near-infrared spectroscopy (fNIRS) is a noninvasive neuroimaging technique to monitor brain activity indirectly. It measures the intensity of near-infrared light diffusely scattered through cortical tissues to quantify changes in oxyhemoglobin concentration (HbO) and deoxyhemoglobin concentration (HbR) with respect to cortical regions.1 – 4 HbO and HbR time series can reflect changes in neurovascular coupling and hence, neuronal activity. fNIRS has been widely used in cognitive,5 – 9 motor skill studies,10 – 15 and brain–computer interface technique.16
- 21 September 2021
- Published by Springer Nature
- p. 467-476
Cardiovascular Disease Risk Improves COVID-19 Patient Outcome Prediction
- 21 September 2021
- Published by Springer Nature
- p. 467-476
Cardiovascular Disease Risk Improves COVID-19 Patient Outcome Prediction
- Open Access
Task-Oriented Low-Dose CT Image Denoising
The extensive use of medical CT has raised a public concern over the radiation dose to the patient. Reducing the radiation dose leads to increased CT image noise and artifacts, which can adversely affect not only the radiologists judgement but also the performance of downstream medical image analysis tasks. Various low-dose CT denoising methods, especially the recent deep learning based approaches, have produced impressive results. However, the existing denoising methods are all downstream-task-agnostic and neglect the diverse needs of the downstream applications. In this paper, we introduce a novel Task-Oriented Denoising Network (TOD-Net) with a task-oriented loss leveraging knowledge from the downstream tasks. Comprehensive empirical analysis shows that the task-oriented loss complements other task agnostic losses by steering the denoiser to enhance the image quality in the task related regions of interest. Such enhancement in turn brings general boosts on the performance of various methods for the downstream task. The presented work may shed light on the future development of context-aware image denoising methods.
Open AccessTask-Oriented Low-Dose CT Image Denoising
The extensive use of medical CT has raised a public concern over the radiation dose to the patient. Reducing the radiation dose leads to increased CT image noise and artifacts, which can adversely affect not only the radiologists judgement but also the performance of downstream medical image analysis tasks. Various low-dose CT denoising methods, especially the recent deep learning based approaches, have produced impressive results. However, the existing denoising methods are all downstream-task-agnostic and neglect the diverse needs of the downstream applications. In this paper, we introduce a novel Task-Oriented Denoising Network (TOD-Net) with a task-oriented loss leveraging knowledge from the downstream tasks. Comprehensive empirical analysis shows that the task-oriented loss complements other task agnostic losses by steering the denoiser to enhance the image quality in the task related regions of interest. Such enhancement in turn brings general boosts on the performance of various methods for the downstream task. The presented work may shed light on the future development of context-aware image denoising methods.
- 11 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 37 (1) , 60-70
Robust Precipitation Bias Correction Through an Ordinal Distribution Autoencoder
Numerical precipitation prediction plays a crucial role in weather forecasting and has broad applications in public services including aviation management and urban disaster early-warning systems. However, numerical weather prediction (NWP) models are often constrained by a systematic bias due to coarse spatial resolution, lack of parameterizations, and limitations of observation and conventional meteorological models, including constrained sample size and long-tail distribution. To address these issues, we present a data-driven deep learning model, named the ordinal distribution autoencoder (ODA), which principally includes a precipitation confidence network and a combinatorial network that contains two blocks, i.e., a denoising autoencoder block and an ordinal distribution regression block. As an expert-free model for bias correction of precipitation, it can effectively correct numerical precipitation prediction based on meteorological data from the European Centre for Medium-Range Weather Forecasts (ECMWF) and SMS-WARMS, an NWP model used in East China. Experiments in the two NWP models demonstrate that, compared with several classical machine-learning algorithms and deep learning models, our proposed ODA generally performs better in bias correction.
- 11 June 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 37 (1) , 60-70
Robust Precipitation Bias Correction Through an Ordinal Distribution Autoencoder
Numerical precipitation prediction plays a crucial role in weather forecasting and has broad applications in public services including aviation management and urban disaster early-warning systems. However, numerical weather prediction (NWP) models are often constrained by a systematic bias due to coarse spatial resolution, lack of parameterizations, and limitations of observation and conventional meteorological models, including constrained sample size and long-tail distribution. To address these issues, we present a data-driven deep learning model, named the ordinal distribution autoencoder (ODA), which principally includes a precipitation confidence network and a combinatorial network that contains two blocks, i.e., a denoising autoencoder block and an ordinal distribution regression block. As an expert-free model for bias correction of precipitation, it can effectively correct numerical precipitation prediction based on meteorological data from the European Centre for Medium-Range Weather Forecasts (ECMWF) and SMS-WARMS, an NWP model used in East China. Experiments in the two NWP models demonstrate that, compared with several classical machine-learning algorithms and deep learning models, our proposed ODA generally performs better in bias correction.