Combining Focus Measures for Three Dimensional Shape Estimation Using Genetic Programming

Three-dimensional (3D) shape reconstruction is a fundamental problem in machine vision applications. Shape from focus (SFF) is one of the passive optical methods for 3D shape recovery, which uses degree of focus as a cue to estimate 3D shape. In this approach, usually a single focus measure operator is applied to measure the focus quality of each pixel in image sequence. However, the applicability of a single focus measure is limited to estimate accurately the depth map for diverse type of real objects. To address this problem, we introduce the development of optimal composite depth (OCD) function through genetic programming (GP) for accurate depth estimation. The OCD function is developed through optimally combining the primary information extracted using one (homogeneous features) or more focus measures (heterogeneous features). The genetically developed composite function is then used to compute the optimal depth map of objects. The performance of this function is investigated using both synthetic and real world image sequences. Experimental results demonstrate that the proposed estimator is more accurate than existing SFF methods. Further, it is found that heterogeneous function is more effective than homogeneous function.

Three-Dimensional Scene Reconstruction

The production of 3D models has been a popular research topic already for a long time, and important progress has been made since the early days. During the last decades, vision systems have established to become the standard and one of the most efficient sensorial assets in industrial and everyday applications. Due to the fact that vision provides several vital attributes, many applications tend to use novel vision systems into domestic, working, industrial, and any other environments. To achieve such goals, a vision system should robustly and effectively reconstruct the 3D surface and the working space. This chapter discusses different methods for capturing the three-dimensional surface of a scene. Geometric approaches to three-dimensional scene reconstruction are generally based on the knowledge of the scene structure from the camera’s internal and external parameters. Another class of methods encompasses the photometric approaches, which evaluate the pixels’ intensity to understand the three-dimensional scene structure. The third and final category of approaches, the so-called real aperture approaches, includes methods that use the physical properties of the visual sensors for image acquisition in order to reproduce the depth information of a scene.

Subject Independent Facial Expression Recognition from 3D Face Models using Deformation Modeling

Most of the works on Facial Expression Recognition (FER) have worked on 2D images or videos. However, researchers are now increasingly utilizing 3D information for FER. As a contribution, this chapter zooms in on 3D based approaches while introducing FER. Prominent works are reviewed briefly, and some of the issues involved in 3D FER are discussed along with the future research directions. In most of the FER approaches, there is a need for having a neutral (expressionless) face of the subject which might not always be practical. This chapter also presents a novel technique of feature extraction which does not require any neutral face of the test subject. A proposition has been verified experimentally that motion of a set of landmark points on the face, in exhibiting a particular facial expression, is similar in different persons. The presented approach shows promising results using Support Vector Machine (SVM) as the classifier.

Stereo-Vision-Based Fire Detection and Suppression Robot for Buildings

A stereo-vision-based fire detection and suppression robot with an intelligent processing algorithm for use in large spaces is proposed in this chapter. The successive processing steps of our real-time algorithm use the motion segmentation algorithm to register the possible position of a fire flame in a video; the real-time algorithm then analyzes the spectral, spatial, and motion orientation characteristics of the fire flame regions from the image sequences of the video. The characterization of a fire flame was carried out by using a heuristic method to determine the potential fire flame candidate region. The fire-fighting robot uses stereo vision generated by means of two calibrated cameras to acquire images of the fire flame and applies the continuously adaptive mean shift (CAMSHIFT) vision-tracking algorithm to provide feedback on the real-time position of the fire flame with a high frame rate. Experimental results showed that the stereo-vision-based mobile robot was able to successfully complete a fire-extinguishing task.

Restoration and Enhancement of Digitally Reconstructed Holographic Images

Holograms can be reconstructed optically or digitally with the use of computers and other related devices. During the reconstruction phase of a hologram by optical or digital methods, some errors may also be introduced that may degrade the quality of obtained hologram, and may lead to a misinterpretation of the holographic image data, which may not be useful for particular application. The basic common errors are zero-order diffraction and speckle noise. These errors have more undesirable effects in digital than in optical holography because the systems of recording and visualization used in the digital holography are extremely sensitive to them or inclusively increase them. The zero-order diffraction can be removed by using high pass filters with low cut-off frequencies and by subtracting the average intensity of all pixels of the hologram image from the original hologram image. Further, the speckle noise introduced during the formation of digital holographic images, which is multiplicative in nature, reduces the image quality, which may not be suitable for specific applications. As the range of applications get broader, demands toward better image quality increases. Hence, the suppression of noise, higher resolution of the reconstructed images, precise parameter adjustment, and faster, more robust algorithms are the essential issues. In this chapter, the various methods available in literature for enhancement and speckle reduction of digital holographic images have been discussed, and a comparative study of results has been presented.

Multi-View Stereo Reconstruction Technique

3D modeling of complex objects is an important task of computer graphics and poses substantial difficulties to traditional synthetic modeling approaches. The multi-view stereo reconstruction technique, which tries to automatically acquire object models from multiple photographs, provides an attractive alternative. The whole reconstruction process of the multi-view stereo technique is introduced in this chapter, from camera calibration and image acquisition to various reconstruction algorithms. The shape from silhouette technique is also introduced since it provides a close shape approximation for many multi-view stereo algorithms. Various multi-view algorithms have been proposed, which can be mainly classified into four classes: 3D volumetric, surface evolution, feature extraction and expansion, and depth map based approaches. This chapter explains the underlying theory and pipeline of each class in detail and analyzes their major properties. Two published benchmarks that are used to qualitatively evaluate multi-view stereo algorithms are presented, along with the benchmark criteria and evaluation results.

Stereo Vision Depth Estimation Methods for Robotic Applications

Vision is undoubtedly the most important sense for humans. Apart from many other low and higher level perception tasks, stereo vision has been proven to provide remarkable results when it comes to depth estimation. As a result, stereo vision is a rather popular and prosperous subject among the computer and machine vision research community. Moreover, the evolution of robotics and the demand for vision-based autonomous behaviors has posed new challenges that need to be tackled. Autonomous operation of robots in real working environments, given limited resources requires effective stereo vision algorithms. This chapter presents suitable depth estimation methods based on stereo vision and discusses potential robotic applications.

3D Scene Capture and Analysis for Intelligent Robotics

The capability of robots to function effectively in the unstructured real world is dominated by the extent to which supporting sensory and computational resources can capture and analyse the 3D working environments within which they are to carry out their tasks. Many device technologies and computational algorithms have been developed over the last 30 years to enable such capabilities. This chapter chronicles a diverse range of such developments, comparing and contrasting them and indicating their various strengths and weaknesses to support intelligent robotic functionality in various domains of application.

The Use of Watermarking in Stereo Imaging

The manipulation and processing of stereo image sequences demand higher costs in memory storage, transmission bandwidth, and computational complexity than of monoscopic images. This chapter investigates scenarios for cost reduction by using reversible watermarking. The basic principle is to embed some data by reversible watermarking instead of either computing or storing/transmitting it. Storage and/or bandwidth are reduced by embedding into one frame of a stereo pair the information needed to recover the other frame. Computational complexity is reduced by embedding the disparity map. The cost of extracting the embedded disparity map is considerably lower than the one of computing it. Experimental results are provided.

Effectiveness of New Technology to Compose Stereoscopic Movies

The most widely known theory of motion sickness and asthenopia are based on the concept of sensory conflict, a disagreement between vergence and visual accommodation while viewing stereoscopic images. Visually induced motion sickness (VIMS) can be measured by using psychological and physiological methods. We quantitatively measured vergence, visual accommodation, head acceleration, and body sway before and during exposure to conventional and new stereoscopic movies. Sickness symptoms appeared with exposure to stereoscopic images. We found that some analytical index for stabilograms increased significantly when the subjects viewed a 3D movie. VIMS could be detected by using these indices. While lateral sway is dependent on the transverse component of head movement while watching the conventional stereoscopic movie, we examine whether this tendency is reduced by Power 3D.