LZLejia Zhang

Lejia Zhang

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Lejia Zhang has a total of 27 co-authors. Here are the top co-authors by number of shared publications:

  • Xiong Wang

    ShanghaiTech University

    8 shared publications
  • Baosheng Wang

    School of Information Science and Technology, ShanghaiTech University, Shanghai, China

    7 shared publications
  • Chenzhe Li

    School of Information Science and Technology, ShanghaiTech University, Shanghai, China

    3 shared publications
  • Zhicheng Wang

    School of Information Science and Technology and the Shanghai Engineering Research Center of Energy Efficient and Custom AI IC, ShanghaiTech University, Shanghai, China

    3 shared publications
  • Chenzhe Li

    School of Information Science and Technology and the Shanghai Engineering Research Center of Energy Efficient and Custom AI IC, ShanghaiTech University, Shanghai, China

    2 shared publications
  • Dantong Liu

    School of Life Science and Technology, ShanghaiTech University, Shanghai, China

    2 shared publications

Latest Publications

  • Deep-Learning-Based Microwave-Induced Thermoacoustic Tomography Applying Realistic Properties of Ultrasound Transducer

    by Lejia Zhang, Qizhi Wang, Simin Zhao, Dantong Liu, Chenzhe Li, Baosheng Wang, Xiong Wang

    Microwave-induced thermoacoustic tomography (MITAT) is a noninvasive hybrid modality that has been widely applied in a bunch of biomedical applications. MITAT integrated with cutting-edge deep learning (DL-MITAT) technique is highly promising for tackling many challenging problems that cannot be efficiently solved by traditional methods. However, the previous MITAT or DL-MITAT works rarely consider the realistic properties of the utilized ultrasound transducer, which can significantly degrade the image quality in some application scenarios. To address this issue, we propose a DL-MITAT technique applying realistic properties of ultrasound transducers (referred to as DL-MITAT-PoT), which is very effective for imaging large or long samples. To be specific, the previous related works simply assume an omnidirectional receiving pattern of the transducer. Instead, we take into account the limited receiving angle of a realistic transducer in the DL-MITAT-PoT technique to improve the imaging performance. To implement this technique, we incorporate the receiving properties of a realistic transducer into the training process of the ResU-Net network. We test the trained network by imaging several 2-mm-diameter long blood-vessel-mimicking samples via both simulations and experiments. The obtained imaging results demonstrate that the proposed DL-MITAT-PoT yields much better imaging quality than its counterpart without considering the property of the transducer. We show that the network trained using complicated vessel sample data is endowed with downward compatibility for samples having less complexity. This technique greatly improves the image quality for dealing with large or long samples and has a promising prospect in many biomedical applications.


  • Investigation of Transducer Distribution in Compressive Thermoacoustic Tomography for Breast Cancer Detection

    by Baosheng Wang, Xiaojuan Zhang, Lejia Zhang, Chenzhe Li, Tao Wu, Xiong Wang

    Microwave-induced thermoacoustic tomography (MITAT) is a novel noninvasive method for biomedical applications like imaging and cancer detection. Recently, compressive sensing has been applied in MITAT, denoted as compressive thermoacoustic tomography (CTAT), can greatly reduce the number of measurements while maintain the image quality. This work reports the study on the effect of distribution of transducer array on image quality in CTAT for the application of breast tumor detection. Different distributions are tested by both simulation and experimental works to compare the quality of imaging. Three breast samples made of pork fat containing tumor-mimicking phantoms are used as the object to evaluate the effect of distribution of transducer array on image quality through the experiment. The results show that the transducer array which distributed evenly on a hemisphere surface is able to achieve an excellent image reconstruction result that is better than other tested types of transducer array distribution. This work gives a faithful guidance on the distribution strategy of transducer array for CTAT in breast tumor detection and other potential applications.


  • Microwave-induced thermoacoustic imaging with a multi-cell AlScN piezoelectric micromachined ultrasonic transducer

    by Yiwei Wang, Lejia Zhang, Baosheng Wang, Junxiang Cai, Yuandong Alex Gu, Liang Lou, Xiong Wang, Tao Wu

    This Letter reports on microwave-induced thermoacoustic imaging utilizing a multi-cell aluminum scandium nitride (AlScN) piezoelectric micromachined ultrasonic transducer (PMUT). Thermoacoustic signals are induced by microwave pulses and then detected by the ultrasonic transducer. It carries biological information, thus enabling noninvasive clinical diagnosis. Our AlScN PMUT device was characterized and experimented to show that it could significantly improve the imaging of the biological sample, decreasing the mean square error by up to 72%, which benefits from a 2.96 times larger detection angle of thermoacoustic signals compared to commercial ultrasonic transducers. The miniaturized and high-performance PMUT devices have the potential to be applied to clinical or personal diagnosis and monitoring equipment.


  • Open Access

    Quantitative Reconstruction of Dielectric Properties Based on Deep-Learning-Enabled Microwave-Induced Thermoacoustic Tomography

    by Zhaoxu Luo, Chenzhe Li, Dantong Liu, Baosheng Wang, Lejia Zhang, Yuexin Ma, Kuiwen Xu, Xiong Wang

    Quantitative reconstruction of dielectric properties has enabled a wealth of biomedical applications. Although traditional microwave imaging and microwave-induced thermoacoustic tomography (MITAT) techniques have been widely explored for quantitative reconstruction, it is still highly challenging for them to deal with biological samples with high permittivity and conductivity. This work leverages deep-learning-enabled MITAT (DL-MITAT) approach to quantitatively reconstruct dielectric properties of biological samples with high quality. We construct a new network structure to separately reconstruct the permittivity and conductivity. By simulation and experimental testing, we demonstrate that the DL-MITAT technique is able to reliably reconstruct inhomogeneous biological samples with tumor, muscle, and fat. The experimental reconstruction error is only 5%. The network exhibits excellent generalization capability in terms of sample’s geometry. This work provides a useful paradigm and alternative way for quantitative reconstruction of dielectric properties and paves the way toward practical applications.


  • Deep-Learning-Enabled Microwave-Induced Thermoacoustic Tomography Based on Sparse Data for Breast Cancer Detection

    by Jiale Zhang, Chenzhe Li, Weichao Jiang, Zhicheng Wang, Lejia Zhang, Xiong Wang

    As a rapidly developing novel electromagnetic imaging technique, microwave-induced thermoacoustic tomography (MITAT) has found many applications and attracted tremendous research interest. Using sparse data to reconstruct images is very challenging for MITAT. This work proposes a novel deep-learning-enabled MITAT (DL-MITAT) modality to address the sparse data reconstruction problem and applies it in breast cancer detection. The applied network is a domain transform network called feature projection network (FPNet) + ResU-Net. Detailed structure and implementation method of the network is described. We conduct both simulation and ex vivo experiments with breast phantoms to test the validity of the DL-MITAT approach. The obtained images given by the trained network exhibit much better quality and have much less artifacts than those obtained by a traditional imaging algorithm. We show that only 15 measurements can still reliably recover an image of the breast tumor for both full-view and limited-view configurations in ex vivo experiments. We also provide detailed discussions on the capability and limitations of the proposed scheme. This work presents a new paradigm for MITAT based on sparse data and can be applied in all related applications of MITAT, including biomedical imaging, nondestructive testing, and therapy guidance.


  • Multiple Back Projection With Impact Factor Algorithm Based on Circular Scanning for Microwave-Induced Thermoacoustic Tomography

    by Qian Song, Zhicheng Wang, Baosheng Wang, Lejia Zhang, Xiong Wang

    Back-projection (BP) algorithm is a common image reconstruction method for microwave-induced thermoacoustic tomography (MITAT). According to the thermoacoustic (TA) signals obtained, which are superimposed with the corresponding time inversion of each pixel in the region of interest (ROI), the BP algorithm can function in a quick response and is suitable for fast real-time MITAT applications. However, high artifacts and background noises caused by incomplete sampling of TA signals seriously affect the image quality. In this paper, a multiple back-projection with impact factor (MBP-IF) algorithm based on circular scanning geometry is proposed to address this problem. Compared with the BP algorithm, the MBP-IF approach calculates not only the initial acoustic pressure of the TA signal of each pixel (i.e., the image obtained by BP), but also the complete acoustic pressure distributions at different time points during the thermoacoustic wave propagation. Then impact factors can be obtained and used to effectively suppress artifacts and eliminate noises. Numerical simulation and phantom experiments prove that the MBP-IF algorithm is advantageous in suppressing image artifacts and noises. Some figures of merit are also obtained to quantify the results. In addition, due to the independence of acoustic pressure distributions at different time points, the MBP-IF can be computed in parallel by GPU, which can be widely used in real-time clinical MITAT system construction.


  • 2-D Noninvasive Temperature Measurement of Biological Samples Based on Compressive Thermoacoustic Tomography

    by Baosheng Wang, Yifei Sun, Chenzhe Li, Zhicheng Wang, Lejia Zhang, Xiong Wang

    There is a great need in temperature monitoring in biomedical applications. The current available temperature measurement techniques have some deficiencies in noninvasiveness, applicability, sensitivity and resolution. Microwave-induced thermoacoustic tomography (MITAT) has found diversified applications in biomedical related disciplines. It has been shown that MITAT can be used to estimate the temperature in a noninvasive manner. However, the applicability of the technique is still limited. This work presents experimental demonstration of two-dimensional (2-D) temperature detection of biological samples based on compressive TAT, referred to as MICTAT. Detailed principle of the method is established. A phantom sample and a biological sample are tested by experiments and 2-D temperature maps are obtained. This work provides a new route of getting 2-D temperature maps of biological samples and may find many applications in biomedical related areas.


  • Microwave-Induced Thermoacoustic Imaging of Small Animals Applying Scanning Orthogonal Polarization Excitation

    by Baosheng Wang, Naping Xiong, Yifei Sun, Lejia Zhang, Chenzhe Li, Jianian Li, Zhicheng Wang, Ziling Chen, Yifeng Zhang, Xiong Wang

    Microwave-induced thermoacoustic tomography (MITAI) has found diversified applications in biomedical related disciplines. This work reports a novel MITAI modality to homogenize the microwave power deposition in the sample and improve the image quality for small animal imaging via scanning orthogonal polarization (SOP) excitation. Although circular polarization excitation has been reported to enhance MITAI, we perform some simulation works using realistic breast phantoms to prove that the homogeneity of microwave power distribution in the phantoms using orthogonal polarization excitation is comparable to that using circular polarization excitation. Compressive sensing is also applied to enhance the time efficiency of the system. Experiments applying a mouse and a frog sample are conducted. The results indicate that the SOP excitation mechanism can completely reveal the structure of the small animals due to improved power homogeneity. The proposed MITAI-SOP technique is superior to circularly polarized antenna (CPA) based modality since the former avoids the deficiency of CPA in bandwidth, power capacity and efficiency. This work presents a practical and easy-to-implement paradigm for high-quality imaging of small animals and big biological samples using MITAI.