Calibrating Laser Three-Dimensional Projection Systems Using Binocular Vision

A laser three-dimensional (3D) projection system is an auxiliary system in intelligent manufacturing. It works with a positioning system in practical applications. This study proposes a calibration method for laser 3D projection systems based on binocular vision. The significance of the binocular vision positioning function for the calibration process was analyzed. Two calibration methods for laser 3D projection systems based on the binocular vision positioning function were proposed. One method involves simplified calculation models and another used data to solve the conversion relationship. The experimental calibration of the projection system was performed using data to directly solve the conversion relationship. The experiment demonstrated the simplicity of the proposed calibration method. The calculation time was less under the 3D laser projection system based on binocular vision. Moreover, the mean calibration error was 0.38 mm at a working distance of 1.8-2.2 m.

5D-fusion imaging for surface shape, polarization, and hyperspectral measurement

We present a five-dimensional (5D) imager that is capable of simultaneous detection of the surface shape, spectral characteristics, and polarization states of macroscopic objects, and straightforwardly fuse collected data into a 5D data set. A polarized module that uses a polarized camera obtains polarized images, while a 3D hyperspectral module reconstructs the target as a 3D point cloud using a fringe projection technique. A liquid-crystal tunable filter is placed in front of the camera of this module to acquire spectral data that can be assigned to corresponding point clouds directly. The two modules are coupled by a dual-path configuration that allows the polarization information to be merged into a comprehensive point cloud with spectral information, generating a new 5D model. The 5D imager shows excellent performance, with a spectral resolution of 10 nm, depth accuracy of 30.7 µm, and imaging time of 8 s. Sample experiments on a toy car with micro scratch defects and a yellowing plant are presented to demonstrate the capabilities of the 5D imager and its potential for use in a broad range of applications, such as industrial manufacturing inspection, plant health monitoring, and biological analysis.

4D dual-mode staring hyperspectral-depth imager for simultaneous spectral sensing and surface shape measurement

A 4D dual-mode staring hyperspectral-depth imager (DSHI), which acquire reflectance spectra, fluorescence spectra, and 3D structural information by combining a staring hyperspectral scanner and a binocular line laser stereo vision system, is introduced. A 405 nm laser line generated by a focal laser line generation module is used for both fluorescence excitation and binocular stereo matching of the irradiated line region. Under the configuration, the two kinds of hyperspectral data collected by the hyperspectral scanner can be merged into the corresponding points in the 3D model, forming a dual-mode 4D model. The DSHI shows excellent performance with spectral resolution of 3 nm, depth accuracy of 26.2 µm. Sample experiments on a fluorescent figurine, real and plastic sunflowers and a clam are presented to demonstrate system's with potential within a broad range of applications such as, e.g., digital documentation, plant phenotyping, and biological analysis.

High-precision four-dimensional hyperspectral imager integrating fluorescence spectral detection and 3D surface shape measurement

A four-dimensional hyperspectral imager (FDHI) that combines fluorescence spectral detection and 3D surface morphology measurement is proposed. The FDHI consists of a hyperspectral line-scanner, a line structured light stereo vision system, and a line laser. The line laser is used as both the excitation light for the fluorescence and the scanning light line for the 3D profiling. At each scanning step, the system collects both fluorescent and 3D spatial data of the irradiated line region, which are fused to 4D data points based on a line mapping relationship between the datasets, and by scanning across the measurement object, a complete 4D dataset is obtained. The FDHI shows excellent performance with spatial and spectral resolution of 26.0 µm and 3 nm, respectively. The reported FDHI system and its applications provide a solution for 4D detection and analysis of fluorescent objects in meters measurement range, with advantage of high integration as two imaging modules sharing a same laser source.

Binocular stereo-microscopy for deforming intact amoeba

A powerful and convenient method for measuring three-dimensional (3D) deformation of moving amoeboid cells will assist the progress of environmental and cytological studies as protists amoebae play a role in the fundamental environmental ecosystem. Here we develop an inexpensive and useful method for measuring 3D deformation of single protists amoeba through binocular microscopy and a newly proposed algorithm of stereo-scopy. From the movies taken from the left and right optical tubes of the binocular microscope, we detect the 3D positions of many intrinsic intracellular vesicles and reconstruct cellular surfaces of amoeboid cells in 3D space. Some observations of sampled behaviors are shown in a single-celled organism of Amoeba proteus. The resultant surface time series is then analyzed to obtain surface velocity, curvature and volume increasing rates of pseudo-pods for characterizing the movements of amoeboid cells. The limitations and errors of this method are also discussed.

Design Considerations for In-Field Measurement of Plant Architecture Traits Using Ground-Based Platforms

This work provides a high-level overview of system design considerations for measuring plant architecture traits in row crops using ground-based, mobile platforms. High-throughput phenotyping technologies are commonly deployed in isolated growth chambers or greenhouses; however, there is a need for field-based systems to measure large quantities of plants exposed to natural climates throughout a growing season. High-throughput methods using ground-based mobile systems collect valuable phenotypic information at higher temporal resolutions compared to manual methods (e.g., handheld calipers and measuring sticks). Additionally, the close proximity to plants when using ground-based systems compared to aerial platforms enables plant phenotyping at the organ level. While there is no single best platform for obtaining ground-based plant measurements across crop varieties with different planting configurations, there are a wide range of off-the-shelf systems and sensors that can be integrated to accommodate varying row widths, plant spacing, plant heights, and plot sizes, in addition to emerging commercially available platforms. This chapter will provide an overview of sensor types suitable for phenotyping plant size and shape, as well as provide guidance for deployment with ground-based systems, including push carts or buggies, modified tractors, and robotic platforms.

SFA-MDEN: Semantic-Feature-Aided Monocular Depth Estimation Network Using Dual Branches

Monocular depth estimation based on unsupervised learning has attracted great attention due to the rising demand for lightweight monocular vision sensors. Inspired by multi-task learning, semantic information has been used to improve the monocular depth estimation models. However, multi-task learning is still limited by multi-type annotations. As far as we know, there are scarcely any large public datasets that provide all the necessary information. Therefore, we propose a novel network architecture Semantic-Feature-Aided Monocular Depth Estimation Network (SFA-MDEN) to extract multi-resolution depth features and semantic features, which are merged and fed into the decoder, with the goal of predicting depth with the support of semantics. Instead of using loss functions to relate the semantics and depth, the fusion of feature maps for semantics and depth is employed to predict the monocular depth. Therefore, two accessible datasets with similar topics for depth estimation and semantic segmentation can meet the requirements of SFA-MDEN for training sets. We explored the performance of the proposed SFA-MDEN with experiments on different datasets, including KITTI, Make3D, and our own dataset BHDE-v1. The experimental results demonstrate that SFA-MDEN achieves competitive accuracy and generalization capacity compared to state-of-the-art methods.

4D surface shape measurement system with high spectral resolution and great depth accuracy

A 4D surface shape measurement system that combines spectral detection and 3D surface morphology measurements is proposed, which can realize high spectral resolution and great depth accuracy (HSDA system). A starring hyperspectral imager system based on a grating generates precise spectral data, while a structured light stereovision system reconstructs target morphology as a 3D point cloud. The systems are coupled using a double light path module, which realize point-to-point correspondence of the systems' image planes. The spectral and 3D coordinate data are fused and transformed into a 4D data set. The HSDA system has excellent performance with a spectral resolution of 3 nm and depth accuracy of 27.5 μm. A range of 4D imaging experiments are presented to demonstrate the capabilities and versatility of the HSDA system, which show that it can be used in broad range of application areas, such as fluorescence detection, face anti-spoofing, physical health state assessment and green plant growth condition monitoring.

Deep Learning for Transient Image Reconstruction from ToF Data

In this work, we propose a novel approach for correcting multi-path interference (MPI) in Time-of-Flight (ToF) cameras by estimating the direct and global components of the incoming light. MPI is an error source linked to the multiple reflections of light inside a scene; each sensor pixel receives information coming from different light paths which generally leads to an overestimation of the depth. We introduce a novel deep learning approach, which estimates the structure of the time-dependent scene impulse response and from it recovers a depth image with a reduced amount of MPI. The model consists of two main blocks: a predictive model that learns a compact encoded representation of the backscattering vector from the noisy input data and a fixed backscattering model which translates the encoded representation into the high dimensional light response. Experimental results on real data show the effectiveness of the proposed approach, which reaches state-of-the-art performances.

Neuromorphic Stereo Vision: A Survey of Bio-Inspired Sensors and Algorithms

Any visual sensor, whether artificial or biological, maps the 3D-world on a 2D-representation. The missing dimension is depth and most species use stereo vision to recover it. Stereo vision implies multiple perspectives and matching, hence it obtains depth from a pair of images. Algorithms for stereo vision are also used prosperously in robotics. Although, biological systems seem to compute disparities effortless, artificial methods suffer from high energy demands and latency. The crucial part is the correspondence problem; finding the matching points of two images. The development of event-based cameras, inspired by the retina, enables the exploitation of an additional physical constraint—time. Due to their asynchronous course of operation, considering the precise occurrence of spikes, Spiking Neural Networks take advantage of this constraint. In this work, we investigate sensors and algorithms for event-based stereo vision leading to more biologically plausible robots. Hereby, we focus mainly on binocular stereo vision.

Flexible Three-Dimensional Reconstruction via Structured-Light-based Visual Positioning and Global Optimization

Three-dimensional (3D) reconstruction using line structured light vision system commonly cooperates with motion restraint devices, such as parallel guide rail push-broom devices. In this study, we propose a visual positioning method to eliminate the motion constraint. An extended orthogonal iteration algorithm for visual positioning is proposed to obtain the precise position of the line structured light binocular camera system during movement. The algorithm uses the information acquired by the binocular camera, and produces a better positioning accuracy than the traditional vision localization algorithm. Furthermore, a global optimization method is proposed to calculate the poses of the camera relative to the world coordinate system at each shooting position. This algorithm effectively reduces the error accumulation and pose drift during visual positioning, and 3D information of the surface can be measured via the proposed free-moving line structured light vision system. The simulation and physical experiments performed herein validate the proposed method and demonstrate the significant improvement in the reconstruction accuracy: when the test distance is 1.5 m, the root mean square error of the point cloud is within 0.5 mm.

FPGA Based Adaptive Rate and Manifold Pattern Projection for Structured Light 3D Camera System

The quality of the captured point cloud and the scanning speed of a structured light 3D camera system depend upon their capability of handling the object surface of a large reflectance variation in the trade-off of the required number of patterns to be projected. In this paper, we propose and implement a flexible embedded framework that is capable of triggering the camera single or multiple times for capturing single or multiple projections within a single camera exposure setting. This allows the 3D camera system to synchronize the camera and projector even for miss-matched frame rates such that the system is capable of projecting different types of patterns for different scan speed applications. This makes the system capturing a high quality of 3D point cloud even for the surface of a large reflectance variation while achieving a high scan speed. The proposed framework is implemented on the Field Programmable Gate Array (FPGA), where the camera trigger is adaptively generated in such a way that the position and the number of triggers are automatically determined according to camera exposure settings. In other words, the projection frequency is adaptive to different scanning applications without altering the architecture. In addition, the proposed framework is unique as it does not require any external memory for storage because pattern pixels are generated in real-time, which minimizes the complexity and size of the application-specific integrated circuit (ASIC) design and implementation.

Optical surgical instrument tracking system based on the principle of stereo vision

Optical tracking systems are widely adopted in surgical navigation. An optical tracking system is designed based on the principle of stereo vision with high-precision and low cost. This system uses optical infrared LEDs that are installed on the surgical instrument as markers and a near-infrared filter is added in front of the Bumblebee2 stereo camera lens to eliminate the interference of ambient light. The algorithm based on the region growing method is designed and used for the marker’s pixel coordinates’ extraction. In this algorithm, the singular points are eliminated and the gray centroid method is applied to solve the pixel coordinate of the marker’s center. Then, the marker’s matching algorithm is adopted and three-dimensional coordinates’ reconstruction is applied to derive the coordinates of the surgical instrument tip in the world coordinate system. In the simulation, the stability, accuracy, rotation tests, and the tracking angle and area range were carried out for a typical surgical instrument and the miniature surgical instrument. The simulation results show that the proposed optical tracking system has high accuracy and stability. It can meet the requirements of surgical navigation.

Quantitative Evaluation of Matching Methods and Validity Measures for Stereo Vision

The authors present a qualitative and quantitative comparison of various similarity measures that form the kernel of common area-based stereo-matching systems. The authors compare classical difference and correlation measures as well as nonparametric measures based on the rank and census transforms for a number of outdoor images. For robotic applications, important considerations include robustness to image defects such as intensity variation and noise, the number of false matches, and computational complexity. In the absence of ground truth data, the authors compare the matching techniques based on the percentage of matches that pass the left-right consistency test. The authors also evaluate the discriminatory power of several match validity measures that are reported in the literature for eliminating false matches and for estimating match confidence. For guidance applications, it is essential to have an estimate of confidence in the three-dimensional points generated by stereo vision. Finally, a new validity measure, the rank constraint, is introduced that is capable of resolving ambiguous matches for rank transform–based matching.

High-resolution, real-time simultaneous 3D surface geometry and temperature measurement

This paper presents a method to simultaneously measure three-dimensional (3D) surface geometry and temperature in real time. Specifically, we developed 1) a holistic approach to calibrate both a structured light system and a thermal camera under exactly the same world coordinate system even though these two sensors do not share the same wavelength; and 2) a computational framework to determine the sub-pixel corresponding temperature for each 3D point as well as discard those occluded points. Since the thermal 2D imaging and 3D visible imaging systems do not share the same spectrum of light, they can perform sensing simultaneously in real time: we developed a hardware system that can achieve real-time 3D geometry and temperature measurement at 26 Hz with 768 × 960 points per frame.

Approaches to three-dimensional reconstruction of plant shoot topology and geometry

There are currently 805million people classified as chronically undernourished, and yet the World's population is still increasing. At the same time, global warming is causing more frequent and severe flooding and drought, thus destroying crops and reducing the amount of land available for agriculture. Recent studies show that without crop climate adaption, crop productivity will deteriorate. With access to 3D models of real plants it is possible to acquire detailed morphological and gross developmental data that can be used to study their ecophysiology, leading to an increase in crop yield and stability across hostile and changing environments. Here we review approaches to the reconstruction of 3D models of plant shoots from image data, consider current applications in plant and crop science, and identify remaining challenges. We conclude that although phenotyping is receiving an increasing amount of attention - particularly from computer vision researchers - and numerous vision approaches have been proposed, it still remains a highly interactive process. An automated system capable of producing 3D models of plants would significantly aid phenotyping practice, increasing accuracy and repeatability of measurements.

Optical Sensors and Methods for Underwater 3D Reconstruction

This paper presents a survey on optical sensors and methods for 3D reconstruction in underwater environments. The techniques to obtain range data have been listed and explained, together with the different sensor hardware that makes them possible. The literature has been reviewed, and a classification has been proposed for the existing solutions. New developments, commercial solutions and previous reviews in this topic have also been gathered and considered.

A quasi-dense matching approach and its calibration application with Internet photos

This paper proposes a quasi-dense matching approach to the automatic acquisition of camera parameters, which is required for recovering 3-D information from 2-D images. An affine transformation-based optimization model and a new matching cost function are used to acquire quasi-dense correspondences with high accuracy in each pair of views. These correspondences can be effectively detected and tracked at the sub-pixel level in multiviews with our neighboring view selection strategy. A two-layer iteration algorithm is proposed to optimize 3-D quasi-dense points and camera parameters. In the inner layer, different optimization strategies based on local photometric consistency and a global objective function are employed to optimize the 3-D quasi-dense points and camera parameters, respectively. In the outer layer, quasi-dense correspondences are resampled to guide a new estimation and optimization process of the camera parameters. We demonstrate the effectiveness of our algorithm with several experiments.

Three-dimensional object motion and velocity estimation using a single computational RGB-D camera

In this paper, a three-dimensional (3D) object moving direction and velocity estimation method is presented using a dual off-axis color-filtered aperture (DCA)-based computational camera. Conventional object tracking methods provided only two-dimensional (2D) states of an object in the image for the target representation. The proposed method estimates depth information in the object region from a single DCA camera that transforms 2D spatial information into 3D model parameters of the object. We also present a calibration method of the DCA camera to estimate the entire set of camera parameters for a practical implementation. Experimental results show that the proposed DCA-based color and depth (RGB-D) camera can calculate the 3D object moving direction and velocity of a randomly moving object in a single-camera framework.

Three-dimensional inline inspection for substrate warpage and ball grid array coplanarity using stereo vision

We present a method for full-field 3D measurement of substrate warpage and ball grid array coplanarity, which is suitable for inline back-end inspection and process monitoring. For evaluating the performance of the proposed system, the linearity between our system and a reference confocal microscope is studied by repeating measurements 35 times with a particular substrate sample (38  mm×28.5  mm). The point-to-point correlation coefficient with 1σ between two methods is 0.968±0.002, and the 2σ difference is 25.15±0.20  μm for warpage measurement. 1σ repeatability of the substrate warpage is 4.2 μm. For BGA coplanarity inspection the bump level correlation coefficient is 0.957±0.001 and the 2σ difference is 28.79±0.14  μm. 1σ repeatability of BGA coplanarity is 3.7 μm. Data acquisition takes about 0.2 s for full field measurements.

Absolute three-dimensional shape measurement using coded fringe patterns without phase unwrapping or projector calibration

This paper presents a novel stereo-phase-based absolute three-dimensional (3D) shape measurement that requires neither phase unwrapping nor projector calibration. This proposed method can be divided into two steps: (1) obtain a coarse disparity map from the quality map; and (2) refine the disparity map using wrapped phase. Fringe patterns are modified to encode the quality map for efficient and accurate stereo matching. Experiments demonstrated that the proposed method could achieve high-quality 3D measurement even with extremely low-quality fringe patterns.

Distance estimation using a single computational camera with dual off-axis color filtered apertures

In this paper, we present a novel computational imaging system using a dual off-axis color filtered aperture (DCA) for distance estimation in a single-camera framework. The DCA consists of two off-axis apertures that are covered by red and cyan color filters. The two apertures generate misaligned color channels in which the amount of misalignment of points in the image plane are a function of the distance from the camera of the corresponding points in the object plane. The primary contribution of this paper is the derivation of a mathematical model of the relationship between the color shifting values and distance of an object from the camera when the camera parameters and the baseline distance between the two apertures in the DCA are given. The proposed computational imaging system can be implemented simply by inserting an appropriately sized DCA into any general optical system. Experimental results show that the DCA camera is able to estimate the distances of objects within a single-camera framework.

A perceptually based comparison of image similarity metrics

The assessment of how well one image matches another forms a critical component both of models of human visual processing and of many image analysis systems. Two of the most commonly used norms for quantifying image similarity are L1 and L2, which are specific instances of the Minkowski metric. However, there is often not a principled reason for selecting one norm over the other. One way to address this problem is by examining whether one metric, better than the other, captures the perceptual notion of image similarity. This can be used to derive inferences regarding similarity criteria the human visual system uses, as well as to evaluate and design metrics for use in image-analysis applications. With this goal, we examined perceptual preferences for images retrieved on the basis of the L1 versus the L2 norm. These images were either small fragments without recognizable content, or larger patterns with recognizable content created by vector quantization. In both conditions the participants showed a small but consistent preference for images matched with the L1 metric. These results suggest that, in the domain of natural images of the kind we have used, the L1 metric may better capture human notions of image similarity.

SYMMETRIC STEREO WITH MULTIPLE WINDOWING

We present a new, efficient stereo algorithm addressing robust disparity estimation in the presence of occlusions. The algorithm is an adaptive, multiwindow scheme using left–right consistency to compute disparity and its associated uncertainty. We demonstrate and discuss performances with both synthetic and real stereo pairs, and show how our results improve on those of closely related techniques for both accuracy and efficiency.

Stereo depth estimation under different camera calibration and alignment errors

Depth estimation is a fundamental issue in computational stereo. To obtain accurate stereo depth estimation, all mechanical parameters with a high precision need to be measured in order to achieve subpixel accuracy and to match features between two different images. This paper investigates accurate depth estimation with different mechanical parameter errors, such as camera calibration and alignment errors, which mainly result from camera lens distortion, camera translation, rotation, pitch, and yaw. For each source of the errors, a model for the error description is presented, and the accurate depth estimation due to this error is quantitatively analyzed. Depth estimation algorithms under an individual error, and with all the errors, are given. Experimental results show that the proposed models can rectify the errors and calculate the accurate depths effectively.

Efficient stereoscopic ranging via stochastic sampling of match quality

We present an efficient method that computes dense stereo correspondences by stochastically sampling match quality values. Nonexhaustive sampling facilitates the use of quality metrics that take unique values at noninteger disparities. Depth estimates are iteratively refined with a stochastic cooperative search by perturbing the estimates, sampling match quality, and reweighting and aggregating the perturbations. The approach gains significant efficiencies when applied to video, where initial estimates are seeded using information from the previous pair in a novel application of the Z-buffering algorithm. This significantly reduces the number of search iterations required. We present a quantitative accuracy evaluation wherein the proposed method outperforms a microcanonical annealing approach by Barnard and a cooperative approach by Zitnick and Kanade , while using fewer match quality evaluations than either. The approach is shown to have more attractive memory usage and scaling than alternatives based on exhaustive sampling.

Differential geometric inference in surface stereo

Many traditional two-view stereo algorithms explicitly or implicitly use the frontal parallel plane assumption when exploiting contextual information since, e.g., the smoothness prior biases toward constant disparity (depth) over a neighborhood. This introduces systematic errors to the matching process for slanted or curved surfaces. These errors are nonnegligible for detailed geometric modeling of natural objects such as a human face. We show how to use contextual information geometrically to avoid such errors. A differential geometric study of smooth surfaces allows contextual information to be encoded in Cartan's moving frame model over local quadratic approximations, providing a framework of geometric consistency for both depth and surface normals; the accuracy of our reconstructions argues for the sufficiency of the approximation. In effect, Cartan's model provides the additional constraint necessary to move beyond the frontal parallel plane assumption in stereo reconstruction. It also suggests how geometry can extend surfaces to account for unmatched points due to partial occlusion.

Vector lifting schemes for stereo image coding

Many research efforts have been devoted to the improvement of stereo image coding techniques for storage or transmission. In this paper, we are mainly interested in lossy-to-lossless coding schemes for stereo images allowing progressive reconstruction. The most commonly used approaches for stereo compression are based on disparity compensation techniques. The basic principle involved in this technique first consists of estimating the disparity map. Then, one image is considered as a reference and the other is predicted in order to generate a residual image. In this paper, we propose a novel approach, based on vector lifting schemes (VLS), which offers the advantage of generating two compact multiresolution representations of the left and the right views. We present two versions of this new scheme. A theoretical analysis of the performance of the considered VLS is also conducted. Experimental results indicate a significant improvement using the proposed structures compared with conventional methods.

Nonlinearities in stereoscopic phase-differencing

Exploiting the quasi-linear relationship between local phase and disparity, phase-differencing registration algorithms provide a fast, powerful means for disparity estimation. Unfortunately, these phase-differencing techniques suffer a significant impediment: phase nonlinearities. In regions of phase nonlinearity, the signals under consideration possess properties that invalidate the use of phase for disparity estimation. This paper uses the amenable properties of Gaussian white noise images to analytically quantify these properties. The improved understanding gained from this analysis enables us to better understand current methodologies for detecting regions of phase instability. Most importantly, we introduce a new, more effective means for identifying these regions based on the second derivative of phase.

On 3-D scene flow and structure recovery from multiview image sequences

Two novel systems computing dense three-dimensional (3-D) scene flow and structure from multiview image sequences are described in this paper. We do not assume rigidity of the scene motion, thus allowing for nonrigid motion in the scene. The first system, integrated model-based system (IMS), assumes that each small local image region is undergoing 3-D affine motion. Non-linear motion model fitting based on both optical flow constraints and stereo constraints is then carried out on each local region in order to simultaneously estimate 3-D motion correspondences and structure. The second system is based on extended gradient-based system (EGS), a natural extension of two-dimensional (2-D) optical flow computation. In this method, a new hierarchical rule-based stereo matching algorithm is first developed to estimate the initial disparity map. Different available constraints under a multiview camera setup are further investigated and utilized in the proposed motion estimation. We use image segmentation information to adopt and maintain the motion and depth discontinuities. Within the framework for EGS, we present two different formulations for 3-D scene flow and structure computation. One formulation assumes that initial disparity map is accurate, while the other does not. Experimental results on both synthetic and real imagery demonstrate the effectiveness of our 3-D motion and structure recovery schemes. Empirical comparison between IMS and EGS is also reported.

On the application of a modified self-organizing neural network to estimate stereo disparity

We propose a modified self-organizing neural network to estimate the disparity map from a stereo pair of images. Novelty consists of the network architecture and of dispensing with the standard assumption of epipolar geometry. Quite distinct from the existing algorithms which, typically, involve area- and/or feature-matching, the network is first initialized to the right image, and then deformed until it is transformed into the left image, or vice versa, this deformation itself being the measure of disparity. Illustrative examples include two classes of stereo pairs: synthetic and natural (including random-dot stereograms and wire frames) and distorted. The latter has one of the following special characteristics: one image is blurred, one image is of a different size, there are salient features like discontinuous depth values at boundaries and surface wrinkles, and there exist occluded and half-occluded regions. While these examples serve, in general, to demonstrate that the technique performs better than many existing algorithms, the above-mentioned stereo pairs (in particular, the last two) bring out some of its limitations, thereby serving as possible motivation for further work.