scholarly journals Semantically Derived Geometric Constraints for MVS Reconstruction of Textureless Areas

2021 ◽  
Vol 13 (6) ◽  
pp. 1053
Author(s):  
Elisavet Konstantina Stathopoulou ◽  
Roberto Battisti ◽  
Dan Cernea ◽  
Fabio Remondino ◽  
Andreas Georgopoulos
Keyword(s):  
Depth Estimation ◽  
Geometric Constraints ◽  
Map Estimation ◽  
Semantic Class ◽  
Novel Approach ◽  
Building Facades ◽  
Indoor Scenes ◽  
Multiview Stereo ◽  
Image Pixels ◽  
Class Labels

Conventional multi-view stereo (MVS) approaches based on photo-consistency measures are generally robust, yet often fail in calculating valid depth pixel estimates in low textured areas of the scene. In this study, a novel approach is proposed to tackle this challenge by leveraging semantic priors into a PatchMatch-based MVS in order to increase confidence and support depth and normal map estimation. Semantic class labels on image pixels are used to impose class-specific geometric constraints during multiview stereo, optimising the depth estimation on weakly supported, textureless areas, commonly present in urban scenarios of building facades, indoor scenes, or aerial datasets. Detecting dominant shapes, e.g., planes, with RANSAC, an adjusted cost function is introduced that combines and weighs both photometric and semantic scores propagating, thus, more accurate depth estimates. Being adaptive, it fills in apparent information gaps and smoothing local roughness in problematic regions while at the same time preserves important details. Experiments on benchmark and custom datasets demonstrate the effectiveness of the presented approach.

scholarly journals The Synthesis of Unpaired Underwater Images for Monocular Underwater Depth Prediction

2021 ◽  
Vol 8 ◽  
Author(s):  
Qi Zhao ◽  
Ziqiang Zheng ◽  
Huimin Zeng ◽  
Zhibin Yu ◽  
Haiyong Zheng ◽  
...  
Keyword(s):  
Structural Information ◽  
Depth Map ◽  
Depth Estimation ◽  
Image Synthesis ◽  
Map Estimation ◽  
Estimation Model ◽  
Depth Prediction ◽  
Underwater Image ◽  
Underwater Vision ◽  
Image Depth

Underwater depth prediction plays an important role in underwater vision research. Because of the complex underwater environment, it is extremely difficult and expensive to obtain underwater datasets with reliable depth annotation. Thus, underwater depth map estimation with a data-driven manner is still a challenging task. To tackle this problem, we propose an end-to-end system including two different modules for underwater image synthesis and underwater depth map estimation, respectively. The former module aims to translate the hazy in-air RGB-D images to multi-style realistic synthetic underwater images while retaining the objects and the structural information of the input images. Then we construct a semi-real RGB-D underwater dataset using the synthesized underwater images and the original corresponding depth maps. We conduct supervised learning to perform depth estimation through the pseudo paired underwater RGB-D images. Comprehensive experiments have demonstrated that the proposed method can generate multiple realistic underwater images with high fidelity, which can be applied to enhance the performance of monocular underwater image depth estimation. Furthermore, the trained depth estimation model can be applied to real underwater image depth map estimation. We will release our codes and experimental setting in https://github.com/ZHAOQIII/UW_depth.

scholarly journals An Automatic Measurement Method for Absolute Depth of Objects in Two Monocular Images Based on SIFT Feature

2017 ◽  
Author(s):  
Lixin He ◽  
Jing Yang ◽  
Bin Kong ◽  
Can Wang
Keyword(s):  
Automatic Measurement ◽  
Depth Estimation ◽  
Depth Information ◽  
Scale Invariant ◽  
Line Segments ◽  
Sift Feature ◽  
Novel Approach ◽  
Straight Line ◽  
Object Depth ◽  
Scale Invariant Feature

It is one of very important and basic problem in compute vision field that recovering depth information of objects from two-dimensional images. In view of the shortcomings of existing methods of depth estimation, a novel approach based on SIFT (the Scale Invariant Feature Transform) is presented in this paper. The approach can estimate the depths of objects in two images which are captured by an un-calibrated ordinary monocular camera. In this approach, above all, the first image is captured. All of the camera parameters remain unchanged, and the second image is acquired after moving the camera a distance d along the optical axis. Then image segmentation and SIFT feature extraction are implemented on the two images separately, and objects in the images are matched. Lastly, an object depth can be computed by the lengths of a pair of straight line segments. In order to ensure that the best appropriate a pair of straight line segments are chose and reduce the computation, the theory of convex hull and the knowledge of triangle similarity are employed. The experimental results show our approach is effective and practical.

scholarly journals Discovering Latent Class Labels for Multi-Label Learning

2020 ◽  
Author(s):  
Jun Huang ◽  
Linchuan Xu ◽  
Jing Wang ◽  
Lei Feng ◽  
Kenji Yamanishi
Keyword(s):  
Large Scale ◽  
Latent Class ◽  
Training Data ◽  
Data Sets ◽  
Robust Learning ◽  
Large Scale Data ◽  
Novel Approach ◽  
Fixed Set ◽  
Class Labels ◽  
Scale Data

Existing multi-label learning (MLL) approaches mainly assume all the labels are observed and construct classification models with a fixed set of target labels (known labels). However, in some real applications, multiple latent labels may exist outside this set and hide in the data, especially for large-scale data sets. Discovering and exploring the latent labels hidden in the data may not only find interesting knowledge but also help us to build a more robust learning model. In this paper, a novel approach named DLCL (i.e., Discovering Latent Class Labels for MLL) is proposed which can not only discover the latent labels in the training data but also predict new instances with the latent and known labels simultaneously. Extensive experiments show a competitive performance of DLCL against other state-of-the-art MLL approaches.

scholarly journals Weakly Supervised Conditional Random Fields Model for Semantic Segmentation with Image Patches

2020 ◽  
Vol 10 (5) ◽  
pp. 1679
Author(s):  
Xinying Xu ◽  
Yujing Xue ◽  
Xiaoxia Han ◽  
Zhe Zhang ◽  
Jun Xie ◽  
...  
Keyword(s):  
Random Fields ◽  
Conditional Random Fields ◽  
Semantic Category ◽  
Semantic Segmentation ◽  
Potential Energy Function ◽  
Semantic Class ◽  
Image Patches ◽  
Benchmark Datasets ◽  
Weakly Supervised ◽  
Class Labels

Image semantic segmentation (ISS) is used to segment an image into regions with differently labeled semantic category. Most of the existing ISS methods are based on fully supervised learning, which requires pixel-level labeling for training the model. As a result, it is often very time-consuming and labor-intensive, yet still subject to manual errors and subjective inconsistency. To tackle such difficulties, a weakly supervised ISS approach is proposed, in which the challenging problem of label inference from image-level to pixel-level will be particularly addressed, using image patches and conditional random fields (CRF). An improved simple linear iterative cluster (SLIC) algorithm is employed to extract superpixels. for image segmentation. Specifically, it generates various numbers of superpixels according to different images, which can be used to guide the process of image patch extraction based on the image-level labeled information. Based on the extracted image patches, the CRF model is constructed for inferring semantic class labels, which uses the potential energy function to map from the image-level to pixel-level image labels. Finally, patch based CRF (PBCRF) model is used to accomplish the weakly supervised ISS. Experiments conducted on two publicly available benchmark datasets, MSRC and PASCAL VOC 2012, have demonstrated that our proposed algorithm can yield very promising results compared to quite a few state-of-the-art ISS methods, including some deep learning-based models.

scholarly journals Depth Prediction without the Sensors: Leveraging Structure for Unsupervised Learning from Monocular Videos

2019 ◽  
Vol 33 ◽  
pp. 8001-8008 ◽  
Author(s):  
Vincent Casser ◽  
Soeren Pirk ◽  
Reza Mahjourian ◽  
Anelia Angelova
Keyword(s):  
Unsupervised Learning ◽  
Moving Objects ◽  
Robot Navigation ◽  
Main Idea ◽  
Object Motion ◽  
Indoor Navigation ◽  
Depth Prediction ◽  
Novel Approach ◽  
Indoor Scenes ◽  
Scene Depth

Learning to predict scene depth from RGB inputs is a challenging task both for indoor and outdoor robot navigation. In this work we address unsupervised learning of scene depth and robot ego-motion where supervision is provided by monocular videos, as cameras are the cheapest, least restrictive and most ubiquitous sensor for robotics. Previous work in unsupervised image-to-depth learning has established strong baselines in the domain. We propose a novel approach which produces higher quality results, is able to model moving objects and is shown to transfer across data domains, e.g. from outdoors to indoor scenes. The main idea is to introduce geometric structure in the learning process, by modeling the scene and the individual objects; camera ego-motion and object motions are learned from monocular videos as input. Furthermore an online refinement method is introduced to adapt learning on the fly to unknown domains. The proposed approach outperforms all state-of-the-art approaches, including those that handle motion e.g. through learned flow. Our results are comparable in quality to the ones which used stereo as supervision and significantly improve depth prediction on scenes and datasets which contain a lot of object motion. The approach is of practical relevance, as it allows transfer across environments, by transferring models trained on data collected for robot navigation in urban scenes to indoor navigation settings. The code associated with this paper can be found at https://sites.google.com/view/struct2depth.

scholarly journals Cloud photogrammetry with dense stereo for fisheye cameras

2016 ◽  
Vol 16 (22) ◽  
pp. 14231-14248 ◽  
Author(s):  
Christoph Beekmans ◽  
Johannes Schneider ◽  
Thomas Läbe ◽  
Martin Lennefer ◽  
Cyrill Stachniss ◽  
...  
Keyword(s):  
Stereo Matching ◽  
Geometric Constraints ◽  
Large Field ◽  
Cloud Base ◽  
Strong Wind ◽  
Real World Data ◽  
First Case ◽  
Novel Approach ◽  
Stereo Reconstruction ◽  
Fisheye Lenses

Abstract. We present a novel approach for dense 3-D cloud reconstruction above an area of 10 × 10 km2 using two hemispheric sky imagers with fisheye lenses in a stereo setup. We examine an epipolar rectification model designed for fisheye cameras, which allows the use of efficient out-of-the-box dense matching algorithms designed for classical pinhole-type cameras to search for correspondence information at every pixel. The resulting dense point cloud allows to recover a detailed and more complete cloud morphology compared to previous approaches that employed sparse feature-based stereo or assumed geometric constraints on the cloud field. Our approach is very efficient and can be fully automated. From the obtained 3-D shapes, cloud dynamics, size, motion, type and spacing can be derived, and used for radiation closure under cloudy conditions, for example. Fisheye lenses follow a different projection function than classical pinhole-type cameras and provide a large field of view with a single image. However, the computation of dense 3-D information is more complicated and standard implementations for dense 3-D stereo reconstruction cannot be easily applied. Together with an appropriate camera calibration, which includes internal camera geometry, global position and orientation of the stereo camera pair, we use the correspondence information from the stereo matching for dense 3-D stereo reconstruction of clouds located around the cameras. We implement and evaluate the proposed approach using real world data and present two case studies. In the first case, we validate the quality and accuracy of the method by comparing the stereo reconstruction of a stratocumulus layer with reflectivity observations measured by a cloud radar and the cloud-base height estimated from a Lidar-ceilometer. The second case analyzes a rapid cumulus evolution in the presence of strong wind shear.

scholarly journals Quantifying Statistical Interdependence by Message Passing on Graphs—Part I: One-Dimensional Point Processes

2009 ◽  
Vol 21 (8) ◽  
pp. 2152-2202 ◽  
Author(s):  
J. Dauwels ◽  
F. Vialatte ◽  
T. Weber ◽  
A. Cichocki
Keyword(s):  
Time Series ◽  
Message Passing ◽  
Point Processes ◽  
Pairwise Alignment ◽  
Surrogate Data ◽  
Map Estimation ◽  
One Dimensional ◽  
Novel Approach ◽  
Average Similarity ◽  
Stochastic Event

We present a novel approach to quantify the statistical interdependence of two time series, referred to as stochastic event synchrony (SES). The first step is to extract “events” from the two given time series. The next step is to try to align events from one time series with events from the other. The better the alignment, the more similar the two time series are considered to be. More precisely, the similarity is quantified by the following parameters: time delay, variance of the timing jitter, fraction of noncoincident events, and average similarity of the aligned events. The pairwise alignment and SES parameters are determined by statistical inference. In particular, the SES parameters are computed by maximum a posteriori (MAP) estimation, and the pairwise alignment is obtained by applying the max-product algorithm. This letter deals with one-dimensional point processes; the extension to multidimensional point processes is considered in a companion letter in this issue. By analyzing surrogate data, we demonstrate that SES is able to quantify both timing precision and event reliability more robustly than classical measures can. As an illustration, neuronal spike data generated by the Morris-Lecar neuron model are considered.

scholarly journals Pixel-based machine learning and image reconstitution for dot-ELISA pathogen serodiagnosis

2020 ◽  
Author(s):  
Cleo Anastassopoulou ◽  
Athanasios Tsakris ◽  
George P. Patrinos ◽  
Yiannis Manoussopoulos
Keyword(s):  
Machine Learning ◽  
False Negative ◽  
Color Discrimination ◽  
Supervised Machine Learning ◽  
Multivariate Logistic Regression Model ◽  
Enzyme Linked Immunosorbent Assay ◽  
Novel Approach ◽  
Image Pixels ◽  
Diagnostic Applications ◽  
Dot Elisa

AbstractSerological methods serve as a direct or indirect means of pathogen infection diagnosis in plant and animal species, including humans. Dot-ELISA (DE) is an inexpensive and sensitive, solid-state version of the microplate enzyme-linked immunosorbent assay, with a broad range of applications in epidemiology. Yet, its applicability is limited by uncertainties in the qualitative output of the assay due to overlapping dot colorations of positive and negative samples, stemming mainly from the inherent color discrimination thresholds of the human eye. Here, we report a novel approach for unambiguous DE output evaluation by applying machine learning-based pattern recognition of image pixels of the blot using an impartial predictive model rather than human judgment. Supervised machine learning was used to train a classifier algorithm through a built multivariate logistic regression model based on the RGB (“Red”, “Green”, “Blue”) pixel attributes of a scanned DE output of samples of known infection status to a model pathogen (Lettuce big-vein associated virus). Based on the trained and cross-validated algorithm, pixel probabilities of unknown samples could be predicted in scanned DE output images which would then be reconstituted by pixels having probabilities above a cutoff that may be selected at will to yield desirable false positive and false negative rates depending on the question at hand, thus allowing for proper dot classification of positive and negative samples and, hence, accurate diagnosis. Potential improvements and diagnostic applications of the proposed versatile method that translates unique pathogen antigens to the universal basic color language are discussed.

scholarly journals LASER-BASED SLAM WITH EFFICIENT OCCUPANCY LIKELIHOOD MAP LEARNING FOR DYNAMIC INDOOR SCENES

2016 ◽  
Vol III-4 ◽  
pp. 119-126 ◽  
Author(s):  
Li Li ◽  
Jian Yao ◽  
Renping Xie ◽  
Jinge Tu ◽  
Chen Feng
Keyword(s):  
Learning Strategy ◽  
Multiple Scales ◽  
Laser Scanner ◽  
Location Based Services ◽  
Indoor Environments ◽  
Map Building ◽  
Scan Matching ◽  
Map Learning ◽  
Novel Approach ◽  
Indoor Scenes

Location-Based Services (LBS) have attracted growing attention in recent years, especially in indoor environments. The fundamental technique of LBS is the map building for unknown environments, this technique also named as simultaneous localization and mapping (SLAM) in robotic society. In this paper, we propose a novel approach for SLAMin dynamic indoor scenes based on a 2D laser scanner mounted on a mobile Unmanned Ground Vehicle (UGV) with the help of the grid-based occupancy likelihood map. Instead of applying scan matching in two adjacent scans, we propose to match current scan with the occupancy likelihood map learned from all previous scans in multiple scales to avoid the accumulation of matching errors. Due to that the acquisition of the points in a scan is sequential but not simultaneous, there unavoidably exists the scan distortion at different extents. To compensate the scan distortion caused by the motion of the UGV, we propose to integrate a velocity of a laser range finder (LRF) into the scan matching optimization framework. Besides, to reduce the effect of dynamic objects such as walking pedestrians often existed in indoor scenes as much as possible, we propose a new occupancy likelihood map learning strategy by increasing or decreasing the probability of each occupancy grid after each scan matching. Experimental results in several challenged indoor scenes demonstrate that our proposed approach is capable of providing high-precision SLAM results.

scholarly journals Challenge Faced in K-Means Clustering for the Choice of Automatic K-Value for Segmentation of Black Sigatoka Disease in Banana Leaves

2020 ◽  
Vol 9 (3) ◽  
pp. 1179-1187
Keyword(s):  
Clustering Algorithm ◽  
Disease Classification ◽  
Machine Learning Techniques ◽  
Bacterial Disease ◽  
Black Sigatoka ◽  
K Value ◽  
Novel Approach ◽  
Class Labels ◽  
Banana Leaves ◽  
Black Sigatoka Disease

Clustering is defined as grouping similar items . The three types of machine learning techniques are supervised, unsupervised and semi-supervised. In unsupervised technique, there are no class labels given to the input data. Clustering is a type of unsupervised learning technique. Recently clustering is applied in many fields such as medicine, agriculture, biology, computers, finance and robotics. Black sigatoka is a bacterial disease occurring commonly in banana plants .The research currently focuses on segmenting the disease area from non-diseased area.The segmentation class training is done via Trainable Weka Segmentation and we also do segmentation using k-means algorithm. In this paper we propose a novel approach for extraction of the black sigatoka diseased area on banana leaves from images using pixel color values and grouping them into their respective clusters accordingly. This is a segmentation cum clustering algorithm. The novel approach has been proposed to overcome the shortfall of k-means clustering when segmenting using automatic value selection for k-means by using silhouette values.Using this novel approach its easy to cluster and segment at the same time. The segmented image from this algorithm can be used in disease classification tasks.

Sign in / Sign up

Export Citation Format

Share Document