Depth Edges Research Articles

A Virtual View Acquisition Technique for Complex Scenes of Monocular Images Based on Layered Depth Images

With the rapid development of stereoscopic display technology, how to generate high-quality virtual view images has become the key in the applications of 3D video, 3D TV and virtual reality. The traditional virtual view rendering technology maps the reference view into the virtual view by means of 3D transformation, but when the background area is occluded by the foreground object, the content of the occluded area cannot be inferred. To solve this problem, we propose a virtual view acquisition technique for complex scenes of monocular images based on a layered depth image (LDI). Firstly, the depth discontinuities of the edge of the occluded area are reasonably grouped by using the multilayer representation of the LDI, and the depth edge of the occluded area is inpainted by the edge inpainting network. Then, the generative adversarial network (GAN) is used to fill the information of color and depth in the occluded area, and the inpainting virtual view is generated. Finally, GAN is used to optimize the color and depth of the virtual view, and the high-quality virtual view is generated. The effectiveness of the proposed method is proved by experiments, and it is also applicable to complex scenes.

DSRNet: Depth Super-Resolution Network guided by blurry depth and clear intensity edges

Although high resolution (HR) depth images are required in many applications such as virtual reality and autonomous navigation, their resolution and quality generated by consumer depth cameras fall short of the requirements. Existing depth upsampling methods focus on extracting multiscale features of HR color image to guide low resolution (LR) depth upsampling, thus causing blurry and inaccurate edges in depth. In this paper, we propose a depth super-resolution (SR) network guided by blurry depth and clear intensity edges, called DSRNet. DSRNet differentiates effective edges from a number of HR edges with the guidance of blurry depth and clear intensity edges. First, we perform global residual estimation based on an encoder–decoder architecture to extract edge structure from HR color image for depth SR. Then, we distinguish effective edges from HR edges in the decoder side with the guidance of LR depth upsampling. To maintain edges for depth SR, we use intensity edge guidance that extracts clear intensity edges from HR image. Finally, we use residual loss to generate accurate high frequency (HF) residual and reconstruct HR depth maps. Experimental results show that DSRNet successfully reconstructs depth edges in SR results as well as outperforms the state-of-the-art methods in terms of visual quality and quantitative measurements.11The proposed model with some test image pairs are available in https://github.com/lanhui-123/DSRNet.

Depth Edge and Structure Optimization-Based End-to-End Self-Supervised Stereo Matching

This paper addresses the challenge of poor cross-domain generalization performance in deep learning methods for stereo matching, particularly when dealing with unseen scenes or disparity maps lacking ground-truth information. To overcome this issue, we propose a self-supervised network called SANet. The network integrates a lightweight algorithm, AANet_Edge, which is based on depth edges optimization. In SANet, we combine AANet_Edge with a novel algorithm called SDCO, which efficiently extracts depth edges using a segment structure and employs a two-layer optimization framework to generate accurate dense disparity maps. These maps are then utilized for pixel-by-pixel supervised training. Furthermore, SANet incorporates multi-scale reconstructed left maps and multi-scale edges-aware modules to learn the structural features of the input image. To evaluate the effectiveness of SANet, comprehensive experiments are conducted on two standard benchmark datasets, namely KITTI 2012 and KITTI 2015. The experimental results demonstrate that SANet produces accurate disparity maps for unseen scenes or limited images and achieves high cross-domain generalization performance.

Depth Information Precise Completion-GAN: A Precisely Guided Method for Completing Ill Regions in Depth Maps

In the depth map obtained through binocular stereo matching, there are many ill regions due to reasons such as lighting or occlusion. These ill regions cannot be accurately obtained due to the lack of information required for matching. Since the completion model based on Gan generates random results, it cannot accurately complete the depth map. Therefore, it is necessary to accurately complete the depth map according to reality. To address this issue, this paper proposes a depth information precise completion GAN (DIPC-GAN) that effectively uses the Guid layer normalization (GuidLN) module to guide the model for precise completion by utilizing depth edges. GuidLN flexibly adjusts the weights of the guiding conditions based on intermediate results, allowing modules to accurately and effectively incorporate the guiding information. The model employs multiscale discriminators to discriminate results of different resolutions at different generator stages, enhancing the generator’s grasp of overall image and detail information. Additionally, this paper proposes Attention-ResBlock, which enables all ResBlocks in each task module of the GAN-based multitask model to focus on their own task by sharing a mask. Even when the ill regions are large, the model can effectively complement the missing details in these regions. Additionally, the multiscale discriminator in the model enhances the generator’s robustness. Finally, the proposed task-specific residual module can effectively focus different subnetworks of a multitask model on their respective tasks. The model has shown good repair results on datasets, including artificial, real, and remote sensing images. The final experimental results showed that the model’s REL and RMSE decreased by 9.3% and 9.7%, respectively, compared to RDFGan.

Structure Flow-Guided Network for Real Depth Super-resolution

Real depth super-resolution (DSR), unlike synthetic settings, is a challenging task due to the structural distortion and the edge noise caused by the natural degradation in real-world low-resolution (LR) depth maps. These defeats result in significant structure inconsistency between the depth map and the RGB guidance, which potentially confuses the RGB-structure guidance and thereby degrades the DSR quality. In this paper, we propose a novel structure flow-guided DSR framework, where a cross-modality flow map is learned to guide the RGB-structure information transferring for precise depth upsampling. Specifically, our framework consists of a cross-modality flow-guided upsampling network (CFUNet) and a flow-enhanced pyramid edge attention network (PEANet). CFUNet contains a trilateral self-attention module combining both the geometric and semantic correlations for reliable cross-modality flow learning. Then, the learned flow maps are combined with the grid-sampling mechanism for coarse high-resolution (HR) depth prediction. PEANet targets at integrating the learned flow map as the edge attention into a pyramid network to hierarchically learn the edge-focused guidance feature for depth edge refinement. Extensive experiments on real and synthetic DSR datasets verify that our approach achieves excellent performance compared to state-of-the-art methods. Our code is available at: https://github.com/Yuanjiayii/DSR-SFG.

Monocular Depth Estimation Network Based on Swin Transformer

Estimating depth from a single image is challenging because a single 2D image may correspond to many different 3D scenes with the same depth. While most deep learning based depth prediction methods extract depth features using small convolutional kernels with small receptive fields, which results in deformed depth edges and inaccurate depth values of distant objects in the depth estimation results. Aiming at this problem, we propose a depth estimation network based on Swin Transformer and the encoder-decoder structure. We construct the encoder using the Swin Transformer network, which can encode long-range spatial dependency and extract features on various scales and across different channels. The decoder of the proposed network is in charge of fusing the features from the encoder by the operations of interpolation, concatenation, and convolution. Experiments on KITTI and NYUv2 datasets show that our proposed network can get more accurate depth edges and depth values than the state-of-the-art methods.

Learning depth via leveraging semantics: Self-supervised monocular depth estimation with both implicit and explicit semantic guidance

Self-supervised monocular depth estimation has shown great success in learning depth using only images for supervision. In this paper, we propose to enhance self-supervised depth estimation with semantics and propose a novel learning scheme, which incorporates both implicit and explicit semantic guidances. Specifically, we propose to relate depth distributions to the semantic category information by proposing a Semantic-aware Spatial Feature Modulation (SSFM) scheme, which implicitly modulates the semantic and depth features in a joint learning framework. The modulation parameters are generated from semantic labels to acquire category-level guidance. Meanwhile, a semantic-guided ranking loss is proposed to explicitly constrain the estimated depth borders using the corresponding segmentation labels. To avoid the impact brought by erroneous segmentation labels, both robust sampling strategy and prediction uncertainty weighting are proposed for the ranking loss. Extensive experimental results show that our method produces high-quality depth maps with semantically consistent depth distributions and accurate depth edges, outperforming the state-of-the-art methods by significant margins.

Improved deep depth estimation for environments with sparse visual cues

Most deep learning-based depth estimation models that learn scene structure self-supervised from monocular video base their estimation on visual cues such as vanishing points. In the established depth estimation benchmarks depicting, for example, street navigation or indoor offices, these cues can be found consistently, which enables neural networks to predict depth maps from single images. In this work, we are addressing the challenge of depth estimation from a real-world bird’s-eye perspective in an industry environment which contains, conditioned by its special geometry, a minimal amount of visual cues and, hence, requires incorporation of the temporal domain for structure from motion estimation. To enable the system to incorporate structure from motion from pixel translation when facing context-sparse, i.e., visual cue sparse, scenery, we propose a novel architecture built upon the structure from motion learner, which uses temporal pairs of jointly unrotated and stacked images for depth prediction. In order to increase the overall performance and to avoid blurred depth edges that lie in between the edges of the two input images, we integrate a geometric consistency loss into our pipeline. We assess the model’s ability to learn structure from motion by introducing a novel industry dataset whose perspective, orthogonal to the floor, contains only minimal visual cues. Through the evaluation with ground truth depth, we show that our proposed method outperforms the state of the art in difficult context-sparse environments.

Adaptive weighted guided image filtering for depth enhancement in shape-from-focus

Existing shape from focus (SFF) techniques cannot preserve depth edges and fine structural details from a sequence of multi-focus images. Moreover, noise in the sequence affects the accuracy of the depth map. In this paper, a novel depth enhancement algorithm for the SFF based on an adaptive weighted guided image filtering (AWGIF) is proposed to address the above issues. The AWGIF is applied to decompose an initial depth map estimated by the traditional SFF into base and detail layers. In order to preserve the edges accurately in the refined depth map, the guidance image is constructed from the sequence, and the coefficient of the AWGIF is utilized to suppress the noise while enhancing the fine depth details. Experiments on real and synthetic objects demonstrate the superiority of our algorithm in terms of anti-noise, and the ability to preserve depth edges and fine structural details w.r.t. existing methods.

Fast Depth Intra Coding Based on Depth Edge Classification Network in 3D-HEVC

As the extension of high efficiency video coding (HEVC) standard, three dimensional-HEVC (3D-HEVC) is the latest 3D video coding standard. 3D-HEVC adopts many complicated coding algorithms to generate additional intermediate views for 3D video representation, which result in extremely high coding complexity. Therefore, this paper proposed a fast depth intra coding approach to reduce the 3D-HEVC complexity, which is based on convolutional neural network (CNN). First, we established a database based on the independent view of the depth map, which includes coding unit (CU) partition data of the depth map. Second, we constructed a depth edge classification CNN (DEC-CNN) framework to classify the edges for the depth map and embedded the network into a 3D-HEVC test platform. Finally, we utilized the pixel value of the binarized depth image to correct the above classification results. The experimental results demonstrated that our approach can reduce the intra coding time by 72.5% on average under negligible degradation of coding performance. This result outperforms the other state-of-the-art methods to reduce the coding complexity of 3D-HEVC.

A novel cell structure‐based disparity estimation for unsupervised stereo matching

It is well known that preserving depth edges is an effective solution for achieving the accurate disparity map in stereo matching, but many state-of-the-art methods do not preserve depth edges well. In order to solve it efficiently, the cell structure containing irregular and regular shape regions is designed to preserve depth edges. Based on the well-designed cell structure, a novel disparity estimation method for stereo matching is proposed, in which a two-layer disparity optimization method is proposed to refine the disparity plane; it includes the front-parallel disparities computation and slanted-surfaces disparity plane refinement. In the framework of front-parallel disparities computation, a tree-based cost aggregation method is presented to make full use of the segmentation information of cells and then performing semi-global cost aggregation. In the framework of slanted-surfaces disparity plane refinement, a new probability model is proposed that employs Bayesian inference for refining disparities in textureless, weak texture and occluded regions. Experimental results show that higher accuracy could be achieved via the proposed method compared with some known state-of-the-art stereo methods on KITTI 2015 and Middlebury dataset, which are the standard benchmarks for testing the stereo matching methods. It can also be indicated that the proposed method can produce accurate disparity map and have good generalization performance.

HLocalExp-CM: confidence map by hierarchical local expansion moves for accurate stereo matching

We present a stereo matching approach referred to as HLocalExp-CM by exploiting the hierarchical local contextual information and a confidence map based on a new grid structure. The proposed approach preserves fine depth edges and extracts accurate disparities in weak texture, textureless, and repeated texture regions. The proposed approach adopts a two-stage optimization strategy. In the framework of first stage, a multiresolution cost aggregation is minimized to reduce the search space of the disparity plane of each pixel. The second stage iteratively optimizes the confidence map and a global energy function to progressively improve the disparity accuracy for each pixel. The confidence map is estimated through classifying the pixels into distinctive and ambiguous ones by computing the decreasing rate of the multiresolution cost aggregation and then performs a spatial propagation and plane refinement for the update of the disparity of each pixel, thereby successfully eliminating the ambiguity of nondistinctive pixels. The global energy function based on a pairwise Markov random field uses cross-scale cost aggregation for taking advantage of context information of objects in different scenarios on local grid regions, which is different from the deep learning technique uses convolution layers extracting the context information. The proposed approach is evaluated on Middlebury benchmark V3, and is ranked first based on “bad 2.0 all metric,” a widely used criterion for the evaluation of stereo images, while the eighth place on “bad 2.0 nonocc metric” (recorded on July 24, 2021).

Deep Multi-Scale Feature Learning for Defocus Blur Estimation

This paper presents an edge-based defocus blur estimation method from a single defocused image. We first distinguish edges that lie at depth discontinuities (called depth edges, for which the blur estimate is ambiguous) from edges that lie at approximately constant depth regions (called pattern edges, for which the blur estimate is well-defined). Then, we estimate the defocus blur amount at pattern edges only, and explore an interpolation scheme based on guided filters that prevents data propagation across the detected depth edges to obtain a dense blur map with well-defined object boundaries. Both tasks (edge classification and blur estimation) are performed by deep convolutional neural networks (CNNs) that share weights to learn meaningful local features from multi-scale patches centered at edge locations. Experiments on naturally defocused images show that the proposed method presents qualitative and quantitative results that outperform state-of-the-art (SOTA) methods, with a good compromise between running time and accuracy.

Intelligent Recognition Method of Athlete Wrong Movement Based on Image Vision

Current methods of human body movement recognition neglect the depth denoising and edge restoration of movement image, which leads to great error in athletes’ wrong movement recognition and poor application intelligence. Therefore, an intelligent recognition method based on image vision for sports athletes’ wrong actions is proposed. The basic principle, structure, and 3D application of computer image vision technology are defined. Capturing the human body image and point cloud data, the three-dimensional dynamic model of sports athletes action is constructed. The color camera including CCD sensor and CMOS sensor is selected to collect the wrong movement image of athlete and provide image data for the recognition of wrong movement. Wavelet transform coefficient and quantization matrix threshold are introduced to denoise the wrong motion images of athletes. Based on this, the feature of sports athlete’s motion contour image is extracted in spatial frequency domain, and the edge of the image is further recovered by Canny operator. Experimental results show that the proposed method can accurately identify the wrong movements of athletes, and there is no redundancy in the recognition results. Image denoising effect is good and less time-consuming and can provide a reliable basis for related fields.

Color-guided optimization model with reliable self-structure priors for depth map restoration

Depth maps captured by Kinect or time of flight (ToF) cameras have an active role in many visual applications. However, a brutal truth is that these depth maps are often contaminated with compound noise, which includes intrinsic noise and missing pixels. In addition, depth maps captured with ToF-based cameras are low in resolution. As these depth maps carry rich and critical information about 3D space, high quality post-processing is crucial for supporting subsequent visual applications. Previous works were proposed via the guiding of the registered color image and bicubic interpolation as an initialization for the up-sampling task, where challenges arose from texture coping and blurry depth discontinuities. Motivated by these challenges, in this paper, we propose a new optimization model depending on the relative structures of both depth and color images for both depth map filtering and up-sampling tasks. In our general model, two self-structure priors for depth and color images are constructed individually and used for the two tasks. For overcoming the texture coping problem, the color-based and depth-based priors are used near the depth edges and at the homogeneous regions respectively. To this end, we further propose a confidence map at every task for managing where every prior is used. Experimental results on both simulated and real datasets for Kinect and ToF cameras demonstrate that the proposed method has a superior performance than benchmarks.

Infrared and Visible Image Fusion Method Based on ResNet in a Nonsubsampled Contourlet Transform Domain

Although the traditional image fusion method can obtain rich image results, obvious artificial noise and artifacts are often present in the resulting image. Fusion algorithms based on neural networks can avoid the shortcomings of traditional methods, but they are more complex and less flexible. In this study, we proposed a fusion method using the deep residual neural network ResNet152, which can not only effectively suppress artificial noise but also preserve the edge details of the image and improve the efficiency of the neural network. The proposed method is characterized by a multiscale transformation of an infrared image and visible light image in the optimized nonsubsampled contourlet transformation domain, and the deep residual neural network ResNet152 is used to extract the deep features of the low-pass component to guide the fusion of the low-pass component. The bandpass component is fused by taking the modulus maximum. This method can fully retain the global features and structural information of the source image in the result image. Compared to existing fusion methods using public test image sets, the experimental results show that on a subjective level, the fusion method creates sharper depth edges and fewer noise artifacts than traditional fusion methods. From an objective perspective, the average value for the results of the evaluation function is greater than that of other fusion methods.

Single Image Depth Estimation Using Edge Extraction Network and Dark Channel Prior

The key to the depth estimation from a single image lies in inferring the distance of various objects without copying texture while maintaining clear object boundaries. In this paper, we propose depth estimation from a single image using edge extraction network and dark channel prior (DCP). We build an edge extraction network based on generative adversarial networks (GANs) to select valid depth edges from a number of edges in an image. We use DCP to generate a transmission map that is able to represent distance from the camera. Transmission map is generated by conducting minimum value filtering on DCP. First, we concatenate the transmission map with the original RGB image to form a tensor, i.e. RGB + T. Second, we generate an initial depth image from the tensor through the generator by inferring depth from stacked residual blocks. Third, we compare the edge map of the initial depth image with that of the input RGB image to select valid depth edges. Both edge maps are generated by the edge extraction network. Finally, we distinguish real and fake on the generated depth image using a discriminator, and enhances the performance of the generator. Various experiments on NYU, Make3D and MPI Sintel datasets demonstrate that the proposed network generates clear edges in depth images as well as outperforms state-of-the-art methods in terms of visual quality and quantitative measurements.

Towards occlusion-aware multifocal displays

The human visual system uses numerous cues for depth perception, including disparity, accommodation, motion parallax and occlusion. It is incumbent upon virtual-reality displays to satisfy these cues to provide an immersive user experience. Multifocal displays, one of the classic approaches to satisfy the accommodation cue, place virtual content at multiple focal planes, each at a different depth. However, the content on focal planes close to the eye do not occlude those farther away; this deteriorates the occlusion cue as well as reduces contrast at depth discontinuities due to leakage of the defocus blur. This paper enables occlusion-aware multifocal displays using a novel ConeTilt operator that provides an additional degree of freedom --- tilting the light cone emitted at each pixel of the display panel. We show that, for scenes with relatively simple occlusion configurations, tilting the light cones provides the same effect as physical occlusion. We demonstrate that ConeTilt can be easily implemented by a phase-only spatial light modulator. Using a lab prototype, we show results that demonstrate the presence of occlusion cues and the increased contrast of the display at depth edges.

Natural statistics of depth edges modulate perceptual stability.

Binocular fusion relies on matching points in the two eyes that correspond to the same physical feature in the world; however, not all world features are binocularly visible. Near depth edges, some regions of a scene are often visible to only one eye (so-called half occlusions). Accurate detection of these monocularly visible regions is likely to be important for stable visual perception. If monocular regions are not detected as such, the visual system may attempt to binocularly fuse non-corresponding points, which can result in unstable percepts. We investigated the hypothesis that the visual system capitalizes on statistical regularities associated with depth edges in natural scenes to aid binocular fusion and facilitate perceptual stability. By sampling from a large set of stereoscopic natural images with co-registered distance information, we found evidence that monocularly visible regions near depth edges primarily result from background occlusions. Accordingly, monocular regions tended to be more visually similar to the adjacent binocularly visible background region than to the adjacent binocularly visible foreground. Consistent with our hypothesis, perceptual experiments showed that perception tended to be more stable when the image properties of the depth edge were statistically more likely given the probability of occurrence in natural scenes (i.e., when monocular regions were more visually similar to the binocular background). The generality of these results was supported by a parametric study with simulated environments. Exploiting regularities in natural environments may allow the visual system to facilitate fusion and perceptual stability when both binocular and monocular regions are visible.

High-quality depth up-sampling via a supervised classification guided MRF model

In this paper, a Supervised Classification assisted Markov Random Field (SC-MRF) model is proposed for generating high-quality up-sampled depth maps. The proposed model aims to reduce depth bleeding and depth confusion artifacts that can be produced at boundary regions of the up-sampled depth maps. In the proposed model, segmentation of low-resolution (LR) depth map is first used to supervise the classification of corresponding high-resolution (HR) color image. With this supervised classification, not only can the depth edges be retained, but redundant textures in the HR color image can be omitted. The classification result is then introduced into the design of a MRF energy function, and the final up-sampled depth map is obtained by optimizing this energy function with the gradient descent algorithm. For simplicity, classical K-means clustering is adopted to segment the LR depth map into several classes, and a feature-based K-nearest neighbour (K-NN) method is utilized for the supervised classification. With the proposed SC-MRF model, interaction between depths of different classes will be strongly suppressed, meaning depth edges are well preserved. Comparisons with the state-of-the-art demonstrate the strong performance of the proposed method both visually and by quantitative evaluation.