Image feature extraction is an essential step in the procedure of image recognition. In this paper, for images features extracting and recognizing, a novel deep neural network called Gaussian–Bernoulli based Convolutional Deep Belief Network (GCDBN) is proposed. The architecture of the proposed GCDBN consists of several convolutional layers based on Gaussian–Bernoulli Restricted Boltzmann Machine. This architecture can take the advantages of Gaussian-Binary Restricted Boltzmann machine and Convolutional Neural Network. Each convolutional layer is followed by a stochastic pooling layer for down-sampling the feature maps. We evaluated our proposed model on several image benchmarks. The experimental results show that our model is more effective for most of images recognition tasks with comparably low computational cost than some of popular methods, which is suggested that our proposed deep network is a potentially applicable method for real-world image recognition.

Learned image compression has attracted considerable interests in recent years. An analysis transform and a synthesis transform, which can be regarded as coupled transforms, are used to encode an image to latent feature and decode the feature after quantization to reconstruct the image. Inspired by the success of invertible neural networks in generative modeling, invertible modules can be used to construct the coupled analysis and synthesis transforms. Considering the noise introduced in the feature quantization invalidates the invertible process, this paper proposes an Approximately Invertible Neural Network (A-INN) framework for learned image compression. It formulates the rate-distortion optimization in lossy image compression when using INN with quantization, which differentiates from using INN for generative modelling. Generally speaking, A-INN can be used as the theoretical foundation for any INN based lossy compression method. Based on this formulation, A-INN with a progressive denoising module (PDM) is developed to effectively reduce the quantization noise in the decoding. Moreover, a Cascaded Feature Recovery Module (CFRM) is designed to learn high-dimensional feature recovery from low-dimensional ones to further reduce the noise in feature channel compression. In addition, a Frequency-enhanced Decomposition and Synthesis Module (FDSM) is developed by explicitly enhancing the high-frequency components in an image to address the loss of highfrequency information inherent in neural network based image compression, thereby enhancing the reconstructed image quality. Extensive experiments demonstrate that the proposed A-INN framework achieves better or comparable compression efficiency than the conventional image compression approach and state-of-the-art learned image compression methods.