Laboratory of Image Informatics

Research

As the advancement in optical microscopy has pushed our understanding about science towards thinner, deeper, and faster, the quality and complexity of large-volume images generated from these state-of-art microscopes preclude conventional visual and manual analysis. Our lab focused on the data-driven image analysis by integrating the cyberinfrastructure ​and the advanced computational tools.

Supported by IU's world-leading computational infrastructure, our research integrates multidisciplinary approaches, including instrumentation, programming, and simulations, to quantitatively characterize images in space and time. With the development of imaging hardware and software, we are particularly interested in the fundamental questions in science and engineering. For example, we use fluorescent imaging system to investigate the spatiotemporal dynamics, mechanics, and landscape of individual keynote molecules or proteins in the live cell. In addition, we also work with other collaborators to investigate the light-matter interaction at nanoscale, seeking for scalable non-conventional light source and novel materials for quantum computing.

 


Imaging Instrumentation: from single molecules to whole animal, from in vitro to in vivo

liu-lab-fig-2.jpg

We have built various optical imaging systems over broad temporal and spatial scales.

  1. The Fluorescence Lifetime Imaging Microscope (FLIM) (left) enables the ultra-fast registration of photons (~25 ps) at a high spatial resolution (~200nm). This home built system has an unique synchronized scanning system to allow two independent image acquisition and pumping.
  2. The single molecule microscope (middle) integrates multiple imaging modalities in a single system. An inverted objective (1.4 N.A.) allows the widefield, Total internal reflection (TIRF), and highly inclined Illumination (HILO) imaging of single molecules with a high temporal resolution. The upright objective enables the light sheet imaging mode, which is capable of obtaining a high-speed volumetric imaging of a live animal (such as zebrafish).
  3. A rencet developed miniscope (right), in collaboration with neuron scientists), aims to monitor the neural activities in freely-moving mouse/rat. The miniscope features centimeter size and extremely low weight (~3g) to allow the real-time imaging of Ca2+ signaling in brain.

 


Computational image analysis: from analytics to informatics

Various computational models have been utilized in our imaging processing/analysis pipeline to automatically recognize and analyze subjects in the image. For example, machine learning and neural networking are used to specifically and correctly segment individual cells in various environments; supervised machine learning is used for single cell phenotyping, and deep learning is used for customized imaging analysis, such as denoise, image registration, and feature extraction.

 


Application 1: How chromatin moves in DNA break/repair?

Motions of DNA breaks are necessary for the onset of genomic translocations driving the initiation, progression and recurrence of certain cancers. Our preliminary data indicate a transient decrease in chromatin mobility during the DNA damage response (DDR), which may protect genomes from deleterious rearrangements. Here, we propose to combine a novel optical platform (light sheet microscope) and in silico models to investigate the mechanisms regulating chromatin motions during the DDR and to address their functional relevance for genome maintenance.

 


Application 2: Force detection in live cells with FRET-based tension sensor

In addition to the genetic and biochemical alternations, mechanic cues are increasingly recognized as key players at all stages throughout the metastasis cascade. However, the lack of appropriate measurement tools prevent us from studying the mechanic alternations in cancer initiation, progression, and spread. This project aims to explore the mechanic landscape of critical motor proteins in live cells. For example, our recent work investigates the force dynamics of focal adhesion of a breast cancer cell in metastasis. Elevated forces in focal adhesion is required for cell migration.

 


Application 3: Investigate the non-classical light source for quantum computing/information science

Scalable single photon or entangled photon source are critical for quantum computing and quantum information science. We have worked with Dr. Vlad Shalaev's group at Purdue to design, fabricate, and characterize the first titanium nitride (TiN) "superlattice", which features hyperbolic dispersion in the visible range and is named hyperbolic metamaterial (HMM). This CMOS-compatible HMM was further used to enhance the single photon emission from single Nitrogen-Vacancy (NV) centers. Also, TiN HMM strengthen the spin contrast of the NV center by more than two times, a step further towards making CMOS-compatible integrated quantum source for superior quantum computers, cryptography and communication technologies. In addition, we also reported the modulated photon statistics and accelerated photon decay rates of the single photon emission tailored by the 2D graphene sheet. We are proposing to develop a graphene-based quantum chip to enable electrical modulation of photon counting rate, lifetime, and statistics of deposited nanodiamond by external gating voltages on a graphene microchip.

People

Jing Liu, Ph.D.

Principal Investigator
Assistant Professor, Physics

urraliu@iupui.edu

View full profile

Hua Li

Postdoctoral Researcher
linhua@iupui.edu

Charles Park

Ph.D. student
cpa1@iupui.edu

Fadil Iqbal

Ph.D. student
fadiqbal@iu.edu

Garrick Chang

Graduate student
changg@iu.edu

Mengdi Zhang

Graduate student
mz10@iu.edu