University of Michigan

Zhaohui Zhong

Xudong Fan

Theodore B Norris

Girish Kulkarni

Jeffrey Allen Fessler

Ayush Pandey

Nanditha Dissanayake

Karthik Reddy

Seunghyun Lee

Chang-Hua Liu

Kyunghoon Lee

You-Chia Chang

Zhengyu Huang

Wenzhe Zang

Image security is becoming an increasingly important issue with the progress of deep learning based image manipulations, such as deep image inpainting and deep fakes. There has been considerable work to date on detecting such image manipulations using better and better algorithms, with little attention being paid to the possible role hardware advances may have for more powerful algorithms. This disclosure proposes to use a focal stack camera as a novel secure imaging device for localizing inpainted regions in manipulated images. Applying convolutional neural network (CNN) methods to focal stack images achieves significantly better detection accuracy compared to single image based detection.

Focal stack camera as secure imaging device and image manipulation detection method

A method of producing uniform multilayer graphene by chemical vapor deposition (CVD) is provided. The method is limited in size only by CVD reaction chamber size and is scalable to produce multilayer graphene films on a wafer scale that have the same number of layers of graphene throughout substantially the entire film. Uniform bilayer graphene may be produced using a method that does not require assembly of independently produced single layer graphene. The method includes a CVD process wherein a reaction gas is flowed in the chamber at a relatively low pressure compared to conventional processes and the temperature in the reaction chamber is thereafter decreased relatively slowly compared to conventional processes. One application for uniform multilayer graphene is transparent conductors. In processes that require multiple transfers of single layer graphene to achieve multilayer graphene structures, the disclosed method can reduce the number of process steps by at least half.

Uniform multilayer graphene by chemical vapor deposition

Electrochemical sensors for the detection of select analytes are provided. The electrochemical sensors include a barrier layer having a substantially uniformed thickness disposed between a sensing layer and an ion exchange membrane. The barrier layer includes a two-dimensional nanomaterial. The barrier layer has a thickness of less than or equal to about 1 nm. The sensing layer has a thickness of less than or equal to about 10 nm. The sensing layer generates ions in response to select analytes. The barrier layer allows the generation ions to pass therethrough and travel into the ion exchange membrane. The barrier layer acts as a physical barrier to contaminants and larger molecules.

Two-dimensional material based ion exchange membrane sensors

An improved sensing method is provided for rapid analyte detection. The method includes: applying an AC excitation signal to the channel region of the transistor; applying an AC drive signal to the transistor; delivering an analyte of interest to a channel region of a transistor; and monitoring a mixing current of the excitation signal and the drive signal through the transistor, where a change in the mixing current is indicative of the concentration of the analyte of interest.

Graphene nanoelectronic heterodyne sensor for rapid and sensitive vapor detection

A photodetector is provided with a thin film double layer heterostructure. The photodetector is comprised of: a substrate; a channel layer of a transistor deposited onto a top surface of the substrate; a source layer of the transistor deposited on the top surface of the substrate; a drain layer of the transistor deposited on the top surface of the substrate, the source layer and the drain layer disposed on opposing sides of the channel layer; a barrier layer deposited onto the channel layer; and a light absorbing layer deposited on the barrier layer. The light absorbing layer is configured to absorb light and, in response to light incident on the light absorbing layer, electrical conductance of the channel layer is changed through hot carrier tunneling from the light absorbing layer to the channel layer.

Photodetectors based on double layer heterostructures

Systems, methods and devices for the efficient photocurrent generation in single- or multi-walled carbon nanotubes, which includes (SWNTs)/poly [3-hexylthiophene-2,5-diyl] (P3HT) hybrid photovoltaics, and exhibit the following features: photocurrent measurement at individual SWNT/P3HT heterojunctions indicate that both semiconducting (s-) and metallic (m-) SWNTs function as excellent hole acceptors; electrical transport and gate voltage dependent photocurrent indicate that P3HT p-dopes both s-SWNT and m-SWNT, and exciton dissociation is driven by a built-in voltage at the heterojunction. Some embodiments include a mmscale SWNT/P3HT bilayer hybrid photovoltaics using horizontally aligned SWNT arrays, which exhibit greater than 90% effective external quantum efficiency, among other things, which advantageously provide carbon nanomaterial based low cost and high efficiency hybrid photovoltaics.

Growing community of inventors

Ann Arbor, MI, United States of America

Zhaohui Zhong