Weihua He

Tsinghua University | TH · Department of Precision Instruments and Mechanical Engineering

Recovering a sequence of latent sharp frames from a motion-blurred image is a challenging task. The bio-inspired event camera, which produces an event stream with high temporal resolution, has been exploited to promote the recovery performance. However, recovering sharp sequences with arbitrary temporal scales has been ignored for a long time. Exis...

The goal of blurry image deblurring and unfolding task is to recover a single sharp frame or a sequence from a blurry one. Recently, its performance is greatly improved with introduction of a bio-inspired visual sensor, event camera. Most existing event-assisted deblurring methods focus on the design of powerful network architectures and effective...

Artificial intelligence (AI) researchers currently believe that the main approach to building more general model problems is the big AI model, where existing neural networks are becoming deeper, larger and wider. We term this the big model with external complexity approach. In this work we argue that there is another approach called small model wit...

Event cameras have gained popularity in depth estimation due to their superior features such as high temporal resolution, low latency, and low power consumption. Spiking Neural Network (SNN) is a promising approach for processing event camera inputs due to its spike-based event-driven nature. However, SNNs face performance degradation when the netw...

Imaging flow cytometry (IFC) has emerged as an innovative tool for cellular heterogeneity investigation, especially from the aspects of cellular composition, structure, and morphology. However, 2D imaging may bring misleading when handling particles with irregular 3D morphology. Current approaches for 3D morphology analysis usually require offline...

Capable of identifying cellular shape, protein localization, and biochemical specificity, joint characterization of cells via fluorescent-and bright-field imaging is a general protocol in imaging flow cytometry (IFC). However, constrained by a narrow imaging field of view (FOV) (typically 105×168) in traditional IFC [1], each channel usually needs...

Reflecting various physiological states and phenotypes of single cells, intrinsic biophysical characteristics (e.g., mechanical and electrical properties) are reliable and important, label‐free biomarkers for characterizing single cells. However, single‐modal mechanical or electrical properties alone are not specific enough to characterize single c...

In the past years, attention-based Transformers have swept across the field of computer vision, starting a new stage of backbones in semantic segmentation. Nevertheless, semantic segmentation under poor light conditions remains an open problem. Moreover, most papers about semantic segmentation work on images produced by commodity frame-based camera...

Event cameras are considered to have great potential for computer vision and robotics applications because of their high temporal resolution and low power consumption characteristics. However, the event stream output from event cameras has asynchronous, sparse characteristics that existing computer vision algorithms cannot handle. Spiking neural ne...

Video frame interpolation is a challenging task due to the ever-changing real-world scene. Previous methods often calculate the bi-directional optical flows and then predict the intermediate optical flows under the linear motion assumptions, leading to isotropic intermediate flow generation. Follow-up research obtained anisotropic adjustment throug...

Generating stylized audio-driven gestures for robots and virtual avatars has attracted increasing considerations recently. Existing methods require style labels (e.g. speaker identities), or complex preprocessing of data to obtain the style control parameters. In this paper, we propose a new end-to-end flow-based model, which can generate audio-dri...

Imaging flow cytometry has emerged as a powerful and important method for non-invasive and high-throughput characterization of single cells. However, sophisticated high-speed camera and heavy computing power necessary in the current frameworks poses a grand challenge in the trade-off between throughput, system complexity, and spatiotemporal range....

Video frame interpolation is a challenging task due to the ever-changing real-world scene. Previous methods often calculate the bi-directional optical flows and then predict the intermediate optical flows under the linear motion assumptions, leading to isotropic intermediate flow generation. Follow-up research obtained anisotropic adjustment throug...

Back Cover In article number 2200325, Wang and coworkers present a camera‐free intrinsic mechanical cytometry (CFIMC) for high‐throughput single‐cell intrinsic mechanical characterization. The strategy of CFIMC adopts a framework that couples the impedance electrodes with the constriction channel spatially, so that the impedance signals contain the...

Mechanical properties of single cells are important label‐free biomarkers normally measured by expensive and complex imaging systems. To unlock this limit and allow mechanical properties comparable across different measurement platforms, camera‐free intrinsic mechanical cytometry (CFIMC) is proposed for on‐the‐fly measurement of two major intrinsic...

Recording fast motion in a high FPS (frame-per-second) requires expensive high-speed cameras. As an alternative, interpolating low-FPS videos from commodity cameras has attracted significant attention. If only low-FPS videos are available, motion assumptions (linear or quadratic) are necessary to infer intermediate frames, which fail to model compl...

Recording fast motion in a high FPS (frame-per-second) requires expensive high-speed cameras. As an alternative, interpolating low-FPS videos from commodity cameras has attracted significant attention. If only low-FPS videos are available, motion assumptions (linear or quadratic) are necessary to infer intermediate frames, which fail to model compl...

Single-cell impedance flow cytometry (IFC) is emerging as a label-free and non-invasive method for characterizing the electrical properties and revealing the sample heterogeneity. At present, most IFC works utilized phenomenological parameters (e.g., impedance amplitude, phase and opacity) to characterize single cells instead of intrinsic biophysic...

This paper presents a novel spatio-temporal domain adaptation (DA) method that is unsupervised and event-driven for dynamic vision sensor (DVS) gesture recognition. This method realizes the transfer between the source domain and the target domain data in both the spatial and temporal dimension without the need of target domain data labels. Specific...

Correction for ‘On-chip stool liquefaction via acoustofluidics’ by Shuaiguo Zhao et al. , Lab Chip , 2019, 19 , 941–947, DOI: 10.1039/C8LC01310A.

Neuromorphic data, recording frameless spike events, have attracted considerable attention for the spatiotemporal information components and the event-driven processing fashion. Spiking neural networks (SNNs) represent a family of event-driven models with spatiotemporal dynamics for neuromorphic computing, which are widely benchmarked on neuromorph...

Neuromorphic data, recording frameless spike events, have attracted considerable attention for the spatiotemporal information components and the event-driven processing fashion. Spiking neural networks (SNNs) represent a family of event-driven models with spatiotemporal dynamics for neuromorphic computing, which are widely benchmarked on neuromorph...

In this study, we proposed a microfluidic device with compact structures integrating multiple modalities for cell capture, pairing, fusion, and culture. The microfluidic device is composed of upper and lower parts. The lower part configured with electrodes and capture wells is used for cell trapping/pairing/fusion, while the upper part configured w...

DOI: 10.1002/elps.201800360 The cover image shows a scene inspired by the fairy tale of Swan Lake. Under the action of a pair of magic bridges and magicians in the boat, the princess and the prince are dancing ballet between the bridges, i.e., rotating in opposite directions. The pair of magic bridges in the scene refer to the parallel electrodes i...

Microfluidic-based portable devices for stool analysis are important for detecting established biomarkers for gastrointestinal disorders and understanding the relationship between gut microbiota imbalances and various health conditions, ranging from digestive disorders to neurodegenerative diseases. However, the challenge of processing stool sample...

This paper describes a highly-efficient integrated microfluidic device for single cell capture, cell pairing, cell fusion, and subsequent on-chip cell couplet culture. However, most of the existing microfluidic devices for cell fusion need to take the fused cells out of the device in order to culture cells or culture cells in situ [1-2]. Integratin...

Different types of cells have different inherent markers which reflect the high specificity of cells. In the work, we present a new microfluidic chip with two opposing fibers and 3D electrodes, which can achieve multi-physical parameter measurement. The device utilizes the principle of optical stretcher to capture and stretch single cells efficient...

Cell rotation is widely required in various fields as an important technique for single cell manipulation. Usually, the electro‐rotational manipulation of single cells by dielectrophoresis (DEP) technologies requires at least three electrodes to generate rotating electric fields which induce cells to rotate. Here, we present a novel microfluidic ch...