Patent
Depth map recovery method
Reads0
Chats0
TLDR
In this article, a depth map recovery method is proposed, comprising of a training set by the depth maps of a large number of various objects; A2, establishing a convolutional neural network (CNN), by using a nuclear separation method, acquiring the parameters of a hidden layer, and training the network structure and adjusting the network weight by using depth maps in the training set; A3, in the output layer of the CNN, establishing an auto-regression model aiming at a possible result, and establishing an evaluation index; and A4, inputting an original depthAbstract:
The invention discloses a depth map recovery method, comprising the following steps of A1, constituting a training set by the depth maps of a large number of various objects; A2, establishing a convolutional neural network (CNN), by using a nuclear separation method, acquiring the parameter of a hidden layer, establishing a convolutional network structure, and training the network structure and adjusting the network weight by using the depth maps in the training set; A3, in the output layer of the CNN, establishing an auto-regression model aiming at a possible result, and establishing an evaluation index; and A4, inputting an original depth map acquired by a depth sensor into the CNN, after denoising and classifying, recovering by an AR model, and if not conforming with requirements, inputting the result map into A2 until the high-quality depth map is acquired or the circulation is ended. According to the depth map recovery method, the image with low resolution and low signal to noise ratio acquired from the depth sensor can be recovered by using the depth convolution network. By using the depth map recovery method, the quality of the depth map can be significantly improved, and meanwhile the method for acquiring the depth map is also simplified.read more
Citations
More filters
Patent
Monocular image depth estimation method based on multi-scale CNN and continuous CRF
Xun Wang,Leqing Zhu,Huiyan Wang +2 more
TL;DR: In this article, a monocular image depth estimation method based on multi-scale CNN and continuous CRF is proposed, where a CRF model is used to calculate single-point potential energy according to an output depth map of DCNN, paired sparse potential energy is calculated according to RGB image, and finally an optimized depth map is deduced using a maximized posterior probability (MAP) algorithm.
Patent
Reconstructing a 3d modeled object
Eloi Mehr,Vincent Guitteny +1 more
TL;DR: In this article, a computer-implemented method for reconstructing a 3D modeled object that represents a real object from a 3-D mesh and measured data representative of the real object is proposed.
Patent
Image depth estimation method based on convolutional neural network
Li Ge,Yu Xiangyu +1 more
TL;DR: In this paper, an image depth estimation method based on a convolutional neural network (CNN) was proposed, where the convolution-deconvolution pair neural network model comprises a plurality of different convolution layers, convolution layer pairs and an activation layer.
Patent
3D reconstruction of a real object from a depth map
TL;DR: In this article, a method for learning a function configured for reconstructing, for a class of real objects, a 3D modeled object that represents an instance of the class from a depth map of the instance.
Patent
Depth sensor noise
TL;DR: In this article, a method for learning a function configured for reconstructing, for a class of real objects, a 3D modeled object that represents an instance of the class from a depth map of the instance is described.
References
More filters
Patent
Image quality testing method based on deep convolutional neural network
Guo Lihua,Li Fudi +1 more
TL;DR: Zhang et al. as discussed by the authors proposed an image quality testing method based on a deep convolutional neural network (DCNN), which consists of firstly establishing a sample set, establishing a deep CNN model, training the CNN model under different initial conditions, connecting the optimal deep CNN models obtained through multiple times of training in parallel, and using the obtained image quality test system for testing images to be tested.
Patent
SAR image terrain classification method based on depth RBF network
Jiao Licheng,Liu Fang,Han Jiamin,Ma Wenping,Ma Jingjing,Wang Shuang,Hou Biao,Li Yangyang,Yang Shuyuan +8 more
TL;DR: Zhang et al. as mentioned in this paper proposed an SAR image terrain classification method based on a depth RBF network, which mainly solves the problem of the prior art that the accuracy of classification is low.
Patent
Facial image normalization method based on self-adaptive multi-column depth model
TL;DR: In this paper, a self-adaptive multi-column depth model was proposed for facial image normalization, and the weight value of each column of depth models was computed in a selfadaptive mode through a nonlinear optimization method, that is, the correction factor of each factor is adjusted in the self-Adaptive mode according to an input image.
Patent
Digital image processing method based on convolutional neural network
Zhang Dai,Qi Hongwen +1 more
TL;DR: In this paper, a digital image processing method based on a convolutional neural network (CNN) is proposed, which consists of three stages: obtaining an original image, extracting the characteristic of the original image through the CNN, and normalizing the new CNN, then performing neuron calculation for obtaining a value of an input signal which is weighted and normalized, and finally obtaining a neuron output activity, thereby obtaining a final image.