Patent•
Image processing method and device, electronic equipment and storage medium
22 May 2020-
TL;DR: In this article, an image processing method and device, including electronic equipment and a storage medium is described, which includes the following steps: carrying out M-level feature extraction on a to-be-processed image to obtain Mlevel first feature maps of the image, wherein the scales of all levels of first features are different, and M is an integer greater than 1; carrying out scale adjustment and fusion on the feature map groups corresponding to the first features at all levels to obtain the M levels of second features.
Abstract: The invention relates to an image processing method and device, electronic equipment and a storage medium. The method comprises the following steps: carrying out M-level feature extraction on a to-be-processed image to obtain M-level first feature maps of the to-be-processed image, wherein the scales of all levels of first feature maps in the M-level first feature maps are different, and M is an integer greater than 1; carrying out scale adjustment and fusion on the feature map groups corresponding to the first feature maps at all levels to obtain M levels of second feature maps, wherein eachfeature map group comprises the first feature map and a first feature map adjacent to the first feature map; and performing target detection on the M-level second feature map to obtain a target detection result of the to-be-processed image. According to the embodiment of the invention, the target detection effect can be improved.
References
More filters
01 Jul 2017
TL;DR: A deep detail network is proposed to directly reduce the mapping range from input to output, which makes the learning process easier and significantly outperforms state-of-the-art methods on both synthetic and real-world images in terms of both qualitative and quantitative measures.
Abstract: We propose a new deep network architecture for removing rain streaks from individual images based on the deep convolutional neural network (CNN). Inspired by the deep residual network (ResNet) that simplifies the learning process by changing the mapping form, we propose a deep detail network to directly reduce the mapping range from input to output, which makes the learning process easier. To further improve the de-rained result, we use a priori image domain knowledge by focusing on high frequency detail during training, which removes background interference and focuses the model on the structure of rain in images. This demonstrates that a deep architecture not only has benefits for high-level vision tasks but also can be used to solve low-level imaging problems. Though we train the network on synthetic data, we find that the learned network generalizes well to real-world test images. Experiments show that the proposed method significantly outperforms state-of-the-art methods on both synthetic and real-world images in terms of both qualitative and quantitative measures. We discuss applications of this structure to denoising and JPEG artifact reduction at the end of the paper.
853 citations
15 Oct 2018
TL;DR: Zhang et al. as discussed by the authors proposed a non-locally enhanced encoder-decoder network, which consists of a pooling indices embedded encoder decoder network to efficiently learn increasingly abstract feature representation for more accurate rain streaks modeling.
Abstract: Single image rain streaks removal has recently witnessed substantial progress due to the development of deep convolutional neural networks. However, existing deep learning based methods either focus on the entrance and exit of the network by decomposing the input image into high and low frequency information and employing residual learning to reduce the mapping range, or focus on the introduction of cascaded learning scheme to decompose the task of rain streaks removal into multi-stages. These methods treat the convolutional neural network as an encapsulated end-to-end mapping module without deepening into the rationality and superiority of neural network design. In this paper, we delve into an effective end-to-end neural network structure for stronger feature expression and spatial correlation learning. Specifically, we propose a non-locally enhanced encoder-decoder network framework, which consists of a pooling indices embedded encoder-decoder network to efficiently learn increasingly abstract feature representation for more accurate rain streaks modeling while perfectly preserving the image detail. The proposed encoder-decoder framework is composed of a series of non-locally enhanced dense blocks that are designed to not only fully exploit hierarchical features from all the convolutional layers but also well capture the long-distance dependencies and structural information. Extensive experiments on synthetic and real datasets demonstrate that the proposed method can effectively remove rain-streaks on rainy image of various densities while well preserving the image details, which achieves significant improvements over the recent state-of-the-art methods.
172 citations
Patent•
11 Jul 2014TL;DR: In this paper, a collection of data that is extremely large can be difficult to search and/or analyze, and a thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products.
Abstract: A collection of data that is extremely large can be difficult to search and/or analyze. Relevance may be dramatically improved by automatically classifying queries and web pages in useful categories, and using these classification scores as relevance features. A thorough approach may require building a large number of classifiers, corresponding to the various types of information, activities, and products. Creation of classifiers and schematizers is provided on large data sets. Exercising the classifiers and schematizers on hundreds of millions of items may expose value that is inherent to the data by adding usable meta-data. Some aspects include active labeling exploration, automatic regularization and cold start, scaling with the number of items and the number of classifiers, active featuring, and segmentation and schematization.
73 citations
01 Nov 2019
TL;DR: Zhang et al. as discussed by the authors proposed a coarse-to-fine single image deraining framework termed Multi-stream Hybrid Deraining Network (shortly, MH-DerainNet), which combines the SSIM loss and perceptual loss to preserve the per-pixel similarity as well as preserving the global structures.
Abstract: Single image deraining task is still a very challenging task due to its ill-posed nature in reality. Recently, researchers have tried to fix this issue by training the CNN-based end-to-end models, but they still cannot extract the negative rain streaks from rainy images precisely, which usually leads to an over de-rained or under de-rained result. To handle this issue, this paper proposes a new coarse-to-fine single image deraining framework termed Multi-stream Hybrid Deraining Network (shortly, MH-DerainNet). To obtain the negative rain streaks during training process more accurately, we present a new module named dual path residual dense block, i.e., Residual path and Dense path. The Residual path is used to reuse com-mon features from the previous layers while the Dense path can explore new features. In addition, to concatenate different scaled features, we also apply the idea of multi-stream with shortcuts between cascaded dual path residual dense block based streams. To obtain more distinct derained images, we combine the SSIM loss and perceptual loss to preserve the per-pixel similarity as well as preserving the global structures so that the deraining result is more accurate. Extensive experi-ments on both synthetic and real rainy images demonstrate that our MH-DerainNet can deliver significant improvements over several recent state-of-the-art methods.
27 citations
TL;DR: Quantitative and qualitative experimental results demonstrate the superiority of the proposed end-to-end deep learning based deraining method compared with several state-of-the-art deraining methods.
Abstract: Rainy images severely degrade visibility and make many computer vision algorithms invalid. Hence, it is necessary to remove rain streaks from a single image. In this paper, we propose a novel end-to-end deep learning based deraining method. Previous methods neglect the correlation between different layers with different receptive fields that loss a lot of important information. To better solve the problem, we develop a multi-level guided residual block that is the basic unit of our network. In this block, we utilize multi-level dilation convolutions to obtain different receptive fields and the layer with smaller receptive fields to guide the learning of larger receptive fields. Moreover, in order to reduce the model sizes, the parameters are shared among all multi-level guided residual blocks. Experiments illustrate that guided learning improves the deraining performance and the shared parameters strategy is also feasible. Quantitative and qualitative experimental results demonstrate the superiority of the proposed method compared with several state-of-the-art deraining methods.
13 citations