Deep Convolutional AutoEncoder-based Lossy Image Compression

Abstract

Image compression has been investigated as a fundamental research topic for many decades. Recently, deep learning has achieved great success in many computer vision tasks, and is gradually being used in image compression. In this paper, we present a lossy image compression architecture, which utilizes the advantages of convolutional autoencoder (CAE) to achieve a high coding efficiency. First, we design a novel CAE architecture to replace the conventional transforms and train this CAE using a rate-distortion loss function. Second, to generate a more energycompact representation, we utilize the principal components analysis (PCA) to rotate the feature maps produced by the CAE, and then apply the quantization and entropy coder to generate the codes. Experimental results demonstrate that our method outperforms traditional image coding algorithms, by achieving a 13.7% BD-rate decrement on the Kodak database images compared to JPEG2000. Besides, our method maintains a moderate complexity similar to JPEG2000.

Deep Joint Source-Channel Coding for Wireless Image Transmission

“…The authors of [8] complement the autoencoder based compression architecture with adversarial loss to achieve realistic reconstructions arXiv:1809.01733v4 [cs.IT] 17 Jun 2019 and improve the visual quality. Cheng et al [9] present a convolutional autoencoder based lossy image compression architecture, which achieves on average a 13.5% rate saving versus JPEG2000 on the Kodak image dataset. The advantage of DL-based methods for lossy compression versus conventional compression algorithms lies in their ability to extract complex features from the training data thanks to their deep architecture, and the fact that their model parameters can be trained efficiently on large datasets through backpropagation.…”

Autoencoder-Based Articulatory-to-Acoustic Mapping for Ultrasound Silent Speech Interfaces

“…a hidden layer, which consists of significantly fewer neurons than the number of input features (or the output layer). Previous studies have shown that this technique can be applied to find relations among the input features [41], for denoising [42], compression [43], and even generating new examples based on the existing ones [44]. Autoencoder neural networks are used for example in image processing [43], [45], audio processing [41] and natural language processing [46].…”

The Application of Image Compression Technology in Library

“…While [9] and [10] utilize the recurrent neural network of LSTM and binary neural layer to replace the quantization and coding process of data. [11] further improves the compression effect by compressing the energy density of data's characteristic pattern.…”

“…While specific to the larger data set, the method can obtain stronger generalization ability due to a variety of training data. Finally, specific to the brand-new digital image resources (such as the virtual reality image shot by Virtual Reality, and 360° panoramic image), the method based on the deep learning can adapt new content and data type [11] more quickly.…”