Image Processing & Applications

Color information is generally considered useful for texture analysis. However, an important category of highly effective texture descriptors - namely rank features - has no obvious extension to color spaces, on which no canonical order is defined. In this work, we explore the use of partial orders in conjunction with rank features. We introduce the rank transform based on product ordering, that generalizes the classic rank transform to RGB space by a combined tally of dominated and non-comparable pixels. Experimental results on nine heterogeneous standard databases confirm that our approach outperforms the standard rank transform and its extension to lexicographic and bit mixing total orders, as well as to the preorders based on the Euclidean distance to a reference color. The low computational complexity and compact codebook size of the transform make it suitable for multi-scale approaches.

Skin appearance is almost universally the object of gender-related expectations and stereotypes. This not with standing, remarkably little work has been done on establishing quantitatively whether skin texture can be used for gender discrimination. We present a detailed analysis of the skin texture of 43 subjects based on two complementary imaging modalities afforded by a visible-light dermoscope and the recently developed Epsilon sensor for capacitive imaging. We consider an array of established texture features in combination with two supervised classification techniques (1-NN and SVM) and a state-of-the-art unsupervised approach (t-SNE). A statistical analysis of the results suggests that skin microtexture carries very little information on gender.

Keywords: Gender recognition , Skin , SVM , Texture

In this paper we introduce Improved Opponent Colour Local Binary Patterns (IOCLBP), a conceptually simple yet effective descriptor for colour texture classification. The method was experimentally validated over eight datasets of colour texture images. The results show that IOCLBP outperformed other LBP variants and was at least as effective as last generation features from Convolutional Neural Networks.

Keywords: Colour texture , Convolutional Neural Networks , Image classification , Local Binary Patterns

This study aims to investigate the feasibility of using time-average holography to verify the integrity of skin tissue samples and detect changes in their mechanical response caused by exposure to thermal perturbations, radiations and mechanical loading. For that purpose, chicken skin samples are put into vibration and the corresponding modes are monitored by means of an optical set up based on time average holographic interferometry. Mode shapes of base samples computed by a parametric finite element model are consistent with experimental data. The holographic set up correctly detects the presence of defects previously included in the sample. Dynamic behavior of skin samples under various conditions of low/high temperature and load or exposed, for different periods of time, to UV radiation and microwaves also is successfully monitored. The reliability and robustness of the proposed approach is tested by performing the holographic observations on a large number of samples under different conditions. Potential benefits that may derive from the use of time-average holography in dermatology and plastic surgery (for example, in the quality control of artificial skin tissues to be implanted) are finally discussed.

Keywords: defects and exposure to heat/radiation/microwaves , modal analysis , Skin tissue , time-average holography

A simple and compact Depth-From-Defocus (DFD) setup, using telecentric illumination and liquid-lens based camera observation, was shown to perform well for 3D shape acquisition over extended measuring range. A further step to ameliorate the system performance is described in this paper. We focused on finding an algorithm to speed up the calibration step of the method, that automatically determines the minimum number of focal lengths to be used in the calibration and measurement procedure. As a result, the calibration is significantly shortened (up to 80% with respect to the original procedure), and the need to manually (and to some extent arbitrarily) select the focal length pairs is overcome. Measurement errors down to 0.73 mm over the measurement depth range of 130 mm, corresponding to 0.55% of the depth range are achieved, in total agreement with the original system.

Keywords: Calibration , Depth From Defocus , Image processing , Liquid lenses , Phase modulation , Shape measurement

We present the in-field use of a recently developed on-line, real time, optical monitoring system for bi-disk rolling contact tests on railway-wheel- and rail- steel disks, based on laser illumination and a high-speed camera. Qualitative, as well as quantitative information can be obtained by the use of specific indexes derived from the images using blob analysis. Monitoring the tests provides information about the initial transients of turning-crest removal, on the steady state reached, and on the dynamics of the surface in between. Blob anisotropy is a key element to draw information of the mechanisms governing this intermediate state.

The experimental characterization of an autofocus algorithm using a liquid lens objective is presented. The objective embeds an electro-wetting based lens whose focal length is voltage controlled. Two sharpness indexes are used to measure the image focus condition in the algorithm allowing a very robust and accurate setting of the focus. The algorithm has been characterized using target images differing both in the contrast and in the spatial frequency along measurement depth range from 70 to 2 mm. both static and dynamic tests were performed revealing the ability of the algorithm to follow rapid variations of the target position.

Keywords: autofocus , Depth-From-Focus , imaging , liquid lenses

A novel depth from defocus (DFD) measurement system is presented, where the extension of the measurement range is performed using an emergent technology based on liquid lenses. A suitable set of different focal lengths, obtained by properly changing the liquid lens supply voltage, provides multiple camera settings without duplicating the system elements or using moving parts. A simple and compact setup, with a single camera/illuminator coaxial assembly, is obtained. The measurement is based on an active DFD technique using modulation measurement profilometry for the estimation of the contrast at each image point as a function of the depth range. Two different measurement methods are proposed, both based on a combination of multiple contrast curves, each derived at a specific focal length. In the first method (intensity contrast method), the depth information is recovered directly from the contrast curves, whereas in the second (differential contrast method), the depth is measured using contrast curve pairs. We obtained a measurement σ0 of 0.55 mm over a depth range of 60 mm with the intensity contrast method (0.92% of the total range) and an σ0 of 0.76 mm over a depth range of 135 mm with the differential contrast method (0.56% of the total range). Thus, the intensity contrast method is within the state-of-the-art DFD systems, whereas the differential contrast method allows, σ0 being almost equal, a remarkable extension of the depth range.

Keywords: Calibration , image edge analysis , image processing , lenses , phase modulation , shape measurement

Wear and Rolling Contact Fatigue (RCF) tests on wheel/rail specimens are important to develop wheels of new materials for improved lifetime and performance, able to operate in harsh environments and at high rolling speeds. We have studied the feasibility of a novel non-invasive all-optical system, based on a high-speed video camera and two laser illumination sources, which is able to continuously monitor the dynamics of the specimens used to test wheel and rail materials, in a laboratory test bench. Surface micro- and macrotopography are monitored using blob analysis and 3D laser triangulation respectively. Blob analysis yields to good discrimination among the specimens, in terms of wear induced surface damage; the 3D measurement, which is characterized by a resolution of 0.033 mm, is able to monitor RCF effects. The system is described with the aid of end-cycle specimens, as well as of intermediate specimens, prior to its installation in the test bench for rolling contact tests.

This work describes a simple, fast, and robust method for identifying, checking and managing the overlapping image keypoints for 3D reconstruction of large objects with numerous geometric singularities and multiple features at different lighting levels. In particular a precision 3D reconstruction of an extensive architecture captured by aerial digital photogrammetry using Unmanned Aerial Vehicles (UAV) is developed. The method was experimentally applied to survey and reconstruct the 'Saraceni' Bridge' at Adrano (Sicily), a valuable example of Roman architecture in brick of historical/cultural interest. The variety of features and different lighting levels required robust self-correlation techniques which would recognise features sometimes even smaller than a pixel in the digital images so as to automatically identify the keypoints necessary for image overlapping and 3D reconstruction. Feature Based Matching (FBM) was used for the low lighting areas like the intrados and the inner arch surfaces, and Area Based Matching (ABM) was used in conjunction to capture the sides and upper surfaces of the bridge. Applying SIFT (Scale Invariant Feature Transform) algorithm during capture helped find distinct features invariant to position, scale and rotation as well as robust for the affinity transformations (changes in scale, rotation, size and position) and lighting variations which are particularly effective in image overlapping. Errors were compared with surveys by total station theodolites, GPS and laser systems. The method can facilitate reconstruction of the most difficult to access parts like the arch intrados and the bridge cavities with high correlation indices.

Keywords: Architectural reconstruction , Area Based Matching , Feature Based Matching , Photogrammetry , SIFT algorithm

Blepharospasm is characterized by bilateral, synchronous, and symmetric involuntary orbicularis oculi muscle spasms leading to partial/total eyelid closure. We proposed a comprehensive method for assessing the severity of blepharospasm cases, based on expert observation of fixed-length video recordings and natural feature detection algorithms. The developed software detects involuntary spasms and blinks in order to evaluate BSP severity. In this work we have considered a new BSP severity scale to realize an objective evaluation of BSP severity.

Shape from shading (SFS) denotes the problem of reconstructing a 3D surface, starting from a single shaded image which represents the surface itself. Minimization techniques are commonly used for solving the SFS problem, where the objective function is a weighted combination of the brightness error, plus one or more terms aiming to obtain a valid solution. We present a regularized quadratic penalty method where quadratic penalization is used to adaptively adjust the smoothing weights, and regularization improves the robustness and reliability of the procedure. A nonmonotone Barzilai–Borwein method is employed to efficiently solve the arising subproblems. Numerical results are provided showing the reliability of the proposed approach.

Keywords: Barzilai–Borwein method , quadratic penalty methods , quadratic regularization , Shape from shading

OBJECTIVES. The purpose of this preliminary in vitro investigation was to evaluate the accuracy of an intraoral scanner in fully edentulous arches rehabilitated with 6 implants; in this clinical condition it is particularly difficult to ensure sufficient accuracy standards. MATERIALS AND METHODS. For this study a reference metal master model was provided: it had 6 implant analogs inserted to simulate a fully rehabilitated edentulous arch with implants. A scanbody made of PEEK (Polyether Ether Ketone) was inserted in each analog and the positions of each scanbody were calibrated by means of a Coordinate Measuring Machine (CMM). The master model, fitted with the scanbodies and a silicone gingiva, was acquired with the True Definition Scanner (TDS). Several series of scans were taken, each of them consisting of 10 close acquisitions, in order to arrange the calibration of the master model and the series of scans with TDS in the same day. From the mesh of each scan, a segmentation was performed to identify known points referable to both plane and cylinder shapes of the scanbodies. This operation was automatically performed by implementing a dedicated algorithm developed in the Rhinoceros 5.0 software. RESULTS. The results herein are related to one set of 10 acquisitions of the master model. Dimensional analysis assessed the effective diameter of the scanbodies measured by TDS and was compared with the reference diameter measured using a CMM, thus obtaining the dimensional error in the body diameter of the scanbody. The values of dimensional errors varied from a minimum of 15 μm to a maximum of about 39 μm, with standard deviations of up to 9 μm. The location analysis allowed to determine the average location error of the scanbodies centers with respect to the reference positions: they were in a range from 14 to 21 μm, with standard deviations of less than 10 μm. CONCLUSIONS. Within the limitations of this in vitro study, the intraoral scanner under investigation proved to be able to scan a 6 implants full arch with a level of accuracy in line with other data reported in scientific literature, or even with smaller error rates. This level of accuracy is capable of fulfilling the accuracy requirements considered clinically acceptable.

Keywords: Full arch digital impressions , Implants , Intraoral scanner , Scanning dimension errors , Scanning position errors

The paper deals with an experimental assessment of the Leap Motion Controller®. This device is able to track the user's hands in a real environment. Due to low-invasiveness and easiness of use, it is promising for the integration in virtual or augmented reality, research and entertainment scenarios. The assessment is performed in a real context using volunteers that were asked to point with the fingertips to a set of predefined locations in space. A specific test rig has been designed and built. It is comprised of a transparent plate supported by adjustable pillars and mounted over the Leap. The data are processed to assess the errors in tracking the five fingertips of the right hand. Results show that the accuracy and precision of the Leap is suitable for robust tracking of the user's hand. The results also unveil that there are preferable zones in which the tracking performance is better.

Keywords: Leap controller , Motion capture , Natural interface , Tracking , User interaction

Introduction of collision models aided surgeons to plan their resection of the femoral and acetabular impinging deformities. Their application on the hip joint is particularly interesting in association with hip arthroscopy because with technique the vision is limited and identifying friction's zones is not easy. Although their large diffusion in the market, they still presents several limits. First, the 3D reconstruction has been used only in the preoperatory phase. Its use inside the operatory room is still not available. Furthermore the actual software lack of a pure automatized program to create the 3D model of the hip. Second, nowadays collision model's software separate and study only the bony part of the hip, erasing from the 3D reconstruction soft tissues. Third, the programs currently developed use statistical data as reference to 3d virtual models and the motion between the hip's bones. They base the motion analysis on the comparison between data extrapolated from the healthy population and data from the patient exam, consequently image processing may be inaccurate. Although according to our opinion collision model are useful even with their actual limits, in this chapter we describe the limits of the actual software and their potential development.

Aim of this study was to validate an inertial system able to detect the loss of ground contact (LOGC) in race-walking through outdoor tests in real training conditions. An inertial sensor was placed at L5/S1 of the vertebral column of a Italian national team athlete to acquire timing measurements of the LOGC. Data were encoded by a well-defined protocol. After a preliminary laboratory study, the athlete performed outdoor-field-tests at different velocities. A specific e-bike with a high-speed camera allowed to acquire a video and to validate sensor measurements. Results indicate that the inertial system can improve the accuracy in detecting the visible LOGC.

Keywords: inertial sensor , loss of contact , outdoor test , Race walking , video analysis

The paper explores the possibility of using low-cost motion capture technologies to automatically evaluate patient's condition concerning his/her walking condition. Two different technologies, optical markerless and inertial, are used to track the gait to be adopted in a doctor's office or at patient's home. The data acquired are elaborated using commercial and in-house developed tools with the aim of creating, in a near future, a simple environment for medical staff and people non highly skilled in IC technology. The paper shows the feasibility of an automatic detection of a set of gait abnormalities affecting people having a lower limb prosthesis. This constitutes a robust support for orthopedic technicians work and foresees the use of such technology for larger surveys and early detection of gait deviations.

Keywords: Gait Analysis , Inertial sensors , Motion Capture , RGB-D sensors

Dedicated imaging methods are among the most important tools of modern computer-aided medical applications. In the last few years, cone beam computed tomography (CBCT) has gained popularity in digital dentistry for 3D imaging of jawbones and teeth. However, the anatomy of a maxillofacial region complicates the assessment of tooth geometry and anatomical location when using standard orthogonal views of the CT data set. In particular, a tooth is defined by a sub-region, which cannot be easily separated from surrounding tissues by only considering pixel grey-intensity values. For this reason, an image enhancement is usually necessary in order to properly segment tooth geometries. In this paper, an anatomy-driven methodology to reconstruct individual 3D tooth anatomies by processing CBCT data is presented. The main concept is to generate a small set of multi-planar reformation images along significant views for each target tooth, driven by the individual anatomical geometry of a specific patient. The reformation images greatly enhance the clearness of the target tooth contours. A set of meaningful 2D tooth contours is extracted and used to automatically model the overall 3D tooth shape through a B-spline representation. The effectiveness of the methodology has been verified by comparing some anatomy-driven reconstructions of anterior and premolar teeth with those obtained by using standard tooth segmentation tools.

Keywords: 3D imaging , B-spline modelling , Cone beam computed tomography , Digital dentistry , Tooth segmentation

One of the most important parameters to be controlled during the production of textile yarns obtained by mixing pre-colored fibers, is the color correspondence between the manufactured yarn and a given reference, usually provided by a designer or a customer. Obtaining yarns from raw pre-colored fibers is a complex manufacturing process entailing a number of steps such as laboratory sampling, color recipe corrections, blowing, carding and spinning. Carding process is the one devoted to transform a “fuzzy mass” of tufted fibers into a regular mass of untwisted fibers, named “tow”. During this process, unfortunately, the correspondence between the color of the tow and the target one cannot be assured, thus leading to yarns whose color differs from the one used for reference. To solve this issue, the main aim of this work is to provide a system able to perform a spectral camera-based real-time measurement of a carded tow, to assess its color correspondence with a reference carded fabric and, at the same time, to monitor the overall quality of the tow during the carding process. Tested against a number of differently colored carded fabrics, the proposed system proved its effectiveness in reliably assessing color correspondence in real-time.

Keywords: Carding process , Color assessment , Spectral camera sensor , Spectrophotometry

The stylistic design of new industrial products often starts from shaded handmade sketches which, usually, need to be converted into 3D digital models by using CAD software packages for the subsequent design phases. This conversion often represents the bottleneck of the whole development process. Shape from Shading (SFS), which attempts to recover the 3D geometry of an object starting from a single shaded representation, is potentially capable of speeding-up the 3D conversion. However, existing approaches prove to be extremely unstable and error prone mainly due to the under-determination of the SFS problem. In order to increase the performance of shading based reconstruction authors propose a step-by-step variational-based approach. In particular, the problem is solved into multiple steps, each one providing the initialization for the following, so that the solution gradually converges towards the final surface. Tested against a set of case studies, the method proved its effectiveness.

Keywords: 3D design , Computer aided design , Numerical optimization , Shape from shading , Stylistic content , Variational approach