Topic
Depth map
About: Depth map is a research topic. Over the lifetime, 8449 publications have been published within this topic receiving 135608 citations.
Papers published on a yearly basis
Papers
More filters
02 Jul 2022
TL;DR: Zhang et al. as discussed by the authors proposed a new depth up-sampling method based on the local edge structure, which can effectively use the pixel correlation and the edge structure between depth map and color map.
Abstract: The current depth up-sampling cannot effectively using the pixel correlation and the edge structure between depth map and color map. In this paper, we proposes a new depth up-sampling method based on the local edge structure. Firstly, the non-robust pixel refinement in the low resolution (LR) depth image is obtained from the LR sobel-based edge map. Then, the structural consistency judgment within the depth map and the color map and the pixels classification is implemented with the guidance of the gradient matrix of the high resolution (HR) color image. Finally, according to the influence between pixels and spatial location constraints, the depth map is refined by the effective depth. Extensive experiments demonstrate that the proposed method outperforms conventional interpolation algorithms and some other edge-based depth up-sampling methods.
TL;DR: In this article , the adaptive sparse depth sampling network is jointly trained with a fusion network of an RGB image and sparse depth, to generate optimal adaptive sampling masks, which can generalize well to many RGB and sparse depths fusion algorithms under a variety of sampling rates.
Abstract: Dense depth map capture is challenging in existing active sparse illumination based depth acquisition techniques, such as LiDAR. Various techniques have been proposed to estimate a dense depth map based on fusion of the sparse depth map measurement with the RGB image. Recent advances in hardware enable adaptive depth measurements resulting in further improvement of the dense depth map estimation. In this paper, we study the topic of estimating dense depth from depth sampling. The adaptive sparse depth sampling network is jointly trained with a fusion network of an RGB image and sparse depth, to generate optimal adaptive sampling masks. Deep learning based superpixel sampling and soft sampling approximation are applied. We show that such adaptive sampling masks can generalize well to many RGB and sparse depth fusion algorithms under a variety of sampling rates (as low as 0.0625%). The proposed adaptive sampling method is fully differentiable and flexible to be trained end-to-end with upstream perception algorithms.
03 Mar 2023
TL;DR: In this paper , the depth estimation results obtained using three different pretrained CNN models which are used as an encoder and decoder are combined to form an ensemble and the average of all the depths predicted by the individual models in the ensemble is considered as the prediction of the ensemble.
Abstract: Estimation of the distance of objects present in the surrounding environment from the visual sensor is the classical research problem in Computer Vision and related areas. The traditional approaches that have been used for depth estimation generally used a pair of images captured through a stereo camera to obtain a disparity map through triangulation method. Later on, researchers have proposed and developed deep learning based methods that can generate the depth map using only a RGB image captured through a monocular camera. However, the recent approaches based on deep convolutional neural networks have shown remarkable development with feasible results. The CNN architecture used for depth estimation consists of two components: Encoder and Decoder. The encoder network is responsible for the extraction of dense features from the input RGB image and the decoder network is used for the upsampling of the encoded depth map. In this paper, we first analyze the depth estimation results obtained using three different pretrained CNN models which are used as an encoder in the depth estimation network, then in order to improve the results further, we leverage the ensemble learning technique, in which these depth estimation models are combined to form an ensemble and the average of all the depths predicted by the individual models in the ensemble is considered as the prediction of the ensemble. The results presented in the paper have shown that the proposed scheme of forming the ensemble of individual encoder-decoder based depth estimation models outperforms the existing benchmark monocular depth estimation methods.
TL;DR: Wang et al. as discussed by the authors proposed a new multi-view stereo reconstruction network integrating depth normal consistency and depth map thinning, which achieved better results in completeness and increased the quality of MVS reconstruction.
Abstract: AbstractTo address the problem of incomplete Multi-view Stereo (MVS) reconstruction, the initial depth and loss function of the depth residual iterative network are investigated, and a new multi-view stereo reconstruction network integrating depth normal consistency and depth map thinning is presented. Firstly, downsampling the input image to create an image pyramid and extracting a feature map from the image pyramid; Then, constructing a cost volume from the 2D feature map, adding the depth normal consistency to the initial cost volume to optimize the depth map. On the DTU data set, the network is tested and compared to traditional reconstruction approaches and MVS networks based on deep learning. The experimental results show that the proposed MVS reconstruction network was produced the better results in completeness and increased the quality of MVS reconstruction.KeywordsNormal-depth consistencyFeature lossCost volumeDepth map refinementMVS
Patent•
04 Jul 2019
TL;DR: In this article, the authors proposed a method for determining a pointing vector pointing from a target point on a target human subject in an image to at least one joint, and determining a positional relationship between the target point and at least 1 pixel point.
Abstract: Embodiments of the present application provide an image processing method and apparatus. In a highly dynamic scenario with a human subject, the present application fully considers the features of limbs or joints of the human subject, thereby improving computation accuracy of a depth map of the human subject. The method comprises: determining a pointing vector pointing from a target point on a target human subject in an image to at least one joint, and determining a positional relationship between the target point and at least one pixel point; and calculating a parallax of the target point according to the pointing vector, the positional relationship, and a parallax of the at least one pixel point, wherein a penalty coefficient of a global energy function of an SGM algorithm is adjusted according to the pointing vector and the positional relationship, and the parallax of the target point is calculated on the basis of the parallax of the at least one pixel point by using the global energy function with the adjusted penalty coefficient.