Ideally, 360° imagery could inherit the deep convolutional neural networks (CNNs) already trained with great success on perspective projection images. However, existing methods to transfer CNNs from perspective to spherical images introduce significant computational costs and/or degradations in accuracy. We present the Kernel Transformer Network (KTN) to efficiently transfer convolution kernels from perspective images to the equirectangular projection of 360° images. Given a source CNN for perspective images as input, the KTN produces a function parameterized by a polar angle and kernel as output. Given a novel 360° image, that function in turn can compute convolutions for arbitrary layers and kernels as would the source CNN on the corresponding tangent plane projections. Distinct from all existing methods, KTNs allow model transfer: the same model can be applied to different source CNNs with the same base architecture. This enables application to multiple recognition tasks without re-training the KTN. Validating our approach with multiple source CNNs and datasets, we show that KTNs improve the state of the art for spherical convolution. KTNs successfully preserve the source CNN’s accuracy, while offering transferability, scalability to typical image resolutions, and, in many cases, a substantially lower memory footprint.

/pdf/kernel-transformer-networks-for-compact-spherical-gd804t1z7w.pdf

Kernel Transformer Networks for Compact Spherical Convolution

Thermoelectric materials and their devices can realize the solid-state energy conversion between thermal and electrical energy, therefore serving as a promising alternative to conventional fossil fuels for energy supply. As...

Advances in Ag2Se-based thermoelectrics from materials to applications

Abstract Exploring new near-room-temperature thermoelectric materials is significant for replacing current high-cost Bi 2 Te 3 . This study highlights the potential of Ag 2 Se for wearable thermoelectric electronics, addressing the trade-off between performance and flexibility. A record-high ZT of 1.27 at 363 K is achieved in Ag 2 Se-based thin films with 3.2 at.% Te doping on Se sites, realized by a new concept of doping-induced orientation engineering. We reveal that Te-doping enhances film uniformity and (00 l )-orientation and in turn carrier mobility by reducing the (00 l ) formation energy, confirmed by solid computational and experimental evidence. The doping simultaneously widens the bandgap, resulting in improved Seebeck coefficients and high power factors, and introduces Te Se point defects to effectively reduce the lattice thermal conductivity. A protective organic-polymer-based composite layer enhances film flexibility, and a rationally designed flexible thermoelectric device achieves an output power density of 1.5 mW cm −2 for wearable power generation under a 20 K temperature difference. 

Flexible power generators by Ag2Se thin films with record-high thermoelectric performance

Advances in bismuth-telluride-based thermoelectric devices: progress and challenges

Abstract Solid‐state bismuth telluride‐based thermoelectric devices enable the generation of electricity from temperature differences and have been commercially applied in various fields. However, in many scenarios, the surface of the heat source is not flat. Therefore, it is crucial to develop flexible thermoelectric materials and devices to efficiently utilize heat sources and expand their applications. Compared with organic thermoelectric materials and devices, inorganic flexible thermoelectric materials and devices have much higher thermoelectric performance and stability. Considering the rapid development in this research field, we carefully summarize the design principles, structures, and thermoelectric properties of inorganic flexible materials and their devices reported in the recent 3 years, including sulfides, selenides, tellurides, and composite materials designed based on these inorganics. The structural designs of flexible thermoelectric devices based on micro‐sized bulk materials are also carefully summarized. Additionally, we overview the mechanical stability and methods for reducing internal resistance for designs of inorganic flexible thermoelectric devices. In the end, we provide outlooks on future research directions for inorganic flexible thermoelectric materials and devices. This review will help guide thermoelectric researchers, beginners, and students. 

Advances in flexible inorganic thermoelectrics

Apparatus and methods for digital video data compression via a scalable, multi-resolution approach. In one embodiment, the video content may be encoded using a multi-resolution and/or multi-quality scalable coding approach that reduces computational and/or energy load on a client device. In one implementation, a low fidelity image is obtained based on a first full resolution image. The low fidelity image may be encoded to obtain a low fidelity bitstream. A second full resolution image may be obtained based on the low fidelity bitstream. A portion of a difference image obtained based on the second full resolution image and the first full resolution may be encoded to obtain a high fidelity bitstream. The low fidelity bitstream and the high fidelity bitstream may be provided to e.g., a receiving device.

Apparatus and methods for video compression using multi-resolution scalable coding

Flexible thermoelectric power generation is a competitive candidate for powering wearable electronic devices and chip-sensor of internet-of-things. Nevertheless, the poor thermoelectric performance of n-type flexible thin film limits its application,...

Significantly (00l)-textured Ag2Se thin films with excellent thermoelectric performance for flexible power applications

Systems and methods for providing imaging content using spatially varying encoding quality. Imaging content may be acquired using spherical lenses (e.g., fisheye). When viewing spherical imaging content, spherical to planar image transformations may be utilized. Such transformations (e.g., equirectangular) may be characterized by spatially varying image distortion. Transformed images may be encoded. Encoding process may be configured based on spatially varying encoding quality. Encoding quality may be configured based on the transformation such that portions of the transformed image characterized by greater distortion may be encoded using lower quality (e.g. greater QP values); portions of the transformed image characterized by lower distortion may be encoded using greater quality (e.g. lower QP values). Such encoding may produce a bitstream characterized by a lower bitrate for a given quality and/or better quality for a given bitrate as compared to an encoding process that may apply a uniform encoding quality.

Systems and methods for compressing video content

Systems and methods for providing video content using spatially adaptive video encoding. Panoramic and/or virtual reality content may be viewed by a client device using a viewport with viewing dimension(s) configured smaller than available dimension(s) of the content. Client device may include a portable media device characterized by given energy and/or computational resources. Video content may be encoded using spatially varying encoding. For image playback, portions of panoramic image may be pre-encoded using multiple quality bands. Pre-encoded image portions, matching the viewport, may be provided and reduce computational and/or energy load on the client device during consumption of panoramic content. Quality distribution may include gradual quality transition area allowing for small movements of the viewport without triggering image re-encoding. Larger movements of the viewport may automatically trigger transition to another spatial encoding distribution.

Systems and methods for spatially adaptive video encoding

Realizing high thermoelectric performance in n-type Bi2Te3 based thin films via post-selenization diffusion

Handle everyday research tasks with reliable, citation-backed results

Your personal Research Agent to handle research tasks with citation-backed results

Popular Tasks used by Researchers

How can I help with your research?

Meet SciSpace

Get more enhanced response by uploading the PDFs you want me to reference.

No relevant PDFs in your library

SciSpace is the AI research assistant for academics. Run systematic literature reviews on 280M+ papers, and write papers with cited sources. Trusted by 1M+ students, PhDs & researchers.

SciSpace | AI for Research

Analyze PDFs

Code & Manuscripts

Funding & Grants

Literature & Patents

Medical & Clinical Data

Systematic Review

Visualize & Present

Web & Data

Build a Google Scholar-like website for your research.

Build a website

Create charts and images for your research

Create a Chart

Write a paper for submission to a journal

Draft a manuscript

Patent Search

Design eye-catching scientific posters in minutes.

Scientific Poster Generation

Systematic Literature Review

One task is running at the moment. Your messages will be shown right after.

Drag and drop or click here to browse

Loved by <highlight>1 million+</highlight> researchers

Extract a list of specific topics and their sources from unstructured text

Topics

Compare and analyze relevant papers that matches with your search

Papers

Get insights from PDFs and bookmarked papers from your library

My library

Recent searches

Try searching for:

Catch AI-generated content in scholarly and non-scholarly content

{ai} Detector

Ai Writer

Get PDF Summaries, highlighted text explanations 

Chat with PDF

Effortlessly create in-text citations and bibliographies in APA and 2,500 other formats

Citation generator

Get explanations, summaries, and answers on academic papers

Ease up your research workflow with {scispace}'s cohort of exciting AI tools

Elevate your academic writing skills and convey your ideas the way you want

Paraphraser

Explore our range of reading and writing tools

Your file is being prepared and should be ready in a few minutes. If it's a large file, it might take a bit longer. You can close this window, and we'll email you the file when it's done.

You have reached a maximum limit of <strong>{limit}</strong> columns in the table. Remove at least <strong>1</strong> column to add or create another one.

Adeel Abbas

Author Tools

Chat about Author