RGB no more: Minimally-decoded JPEG Vision Transformers
Most neural networks for computer vision are designed to infer using RGB images. However,
these RGB images are commonly encoded in JPEG before saving to disk; decoding them …
these RGB images are commonly encoded in JPEG before saving to disk; decoding them …
Deep learning in frequency domain for inverse identification of nonhomogeneous material properties
The inverse identification of nonhomogeneous material properties from measured
displacement/strain fields, especially when noise exists, is crucial for both engineering and …
displacement/strain fields, especially when noise exists, is crucial for both engineering and …
Harmonic networks for image classification
Convolutional neural networks (CNNs) learn filters in order to capture local correlation
patterns in feature space. In contrast, in this paper we propose harmonic blocks that produce …
patterns in feature space. In contrast, in this paper we propose harmonic blocks that produce …
Quannet: Joint image compression and classification over channels with limited bandwidth
The performance of cloud based image classification depends critically on its allocated
bandwidth. Traditional data compression methods can negatively impact classification …
bandwidth. Traditional data compression methods can negatively impact classification …
Improving the accuracy-latency trade-off of edge-cloud computation offloading for deep learning services
Offloading tasks to the edge or the Cloud has the potential to improve accuracy of
classification and detection tasks as more powerful hardware and machine learning models …
classification and detection tasks as more powerful hardware and machine learning models …
Orthogonal features based EEG signals denoising using fractional and compressed one-dimensional CNN AutoEncoder
This paper presents a fractional one-dimensional convolutional neural network (CNN)
autoencoder for denoising the Electroencephalogram (EEG) signals which often get …
autoencoder for denoising the Electroencephalogram (EEG) signals which often get …
Dct-based fast spectral convolution for deep convolutional neural networks
Y Xu, H Nakayama - 2021 International Joint Conference on …, 2021 - ieeexplore.ieee.org
Spectral representations have been introduced into deep convolutional neural networks
(CNNs) mainly for accelerating convolutions and mitigating information loss. However …
(CNNs) mainly for accelerating convolutions and mitigating information loss. However …
Harmonic networks with limited training samples
Convolutional neural networks (CNNs) are very popular nowadays for image processing.
CNNs allow one to learn optimal filters in a (mostly) supervised machine learning context …
CNNs allow one to learn optimal filters in a (mostly) supervised machine learning context …
Improving deep learning classification of jpeg2000 images over bandlimited networks
LD Chamain, Z Ding - ICASSP 2020-2020 IEEE International …, 2020 - ieeexplore.ieee.org
JPEG2000 (j2k) is a highly popular format for image and video compression. It plays a major
role in the rapidly growing applications of cloud based image classification. Considering …
role in the rapidly growing applications of cloud based image classification. Considering …
Training method and apparatus for neural network for image recognition
L Chen, S Wang, W Fan, J Sun, N Satoshi - US Patent 10,296,813, 2019 - Google Patents
A training method and a training apparatus for a neutral network for image recognition are
provided. The method includes: representing a sample image as a point set in a high …
provided. The method includes: representing a sample image as a point set in a high …