Image Processing & Applications

The present work introduces an emotional tracking system to monitor Shopping Experience at different touchpoints in a retail store, based on the elaboration of the information extracted from biometric data and facial expressions. A preliminary test has been carried out to determine the system effectiveness in a real context regarding to emotion detection and customers' sex, age and ethnicity discrimination. To this end, information provided by the system have been compare with the results of a traditional video analysis. Results suggest that the proposed system can be effectively used to support the analysis of customer experience in a retail context.

Keywords: customer experience , emotion analysis , emotion tracking , face recognition , facial expression , shopping experience

The present work introduces an emotional tracking system to monitor Shopping Experience at different touchpoints in a store, based on the elaboration of the information extracted from biometric data and facial expressions. Preliminary tests suggest that the proposed system can be effectively used in a retail context.

Keywords: Customer experience , Emotion analysis , Emotion tracking , Shopping experience

This research aims to develop a system that examines and reacts to the changing behaviors and emotions of individuals in order to improve their shopping experience. The system is able to track emotions in real time at different touchpoints in a store and control a set of networked devices to configure the sensing space and all provided services responsive to the cus-tomers' needs. This paper describes the general approach adopted to design the overall system and illustrates in detail the module prototyped to understand the users' emotions through the analysis of facial expressions.

Keywords: Context-aware computing , Emotion recognition , Methods for CX , Shopping experience

The visualization and analysis of mosaics and pavements are often compromised by their large sizes, which do not enable the observer to perceive their whole arrangement or to focus on details placed in farthest areas from its boundaries. Moreover, the usual precarious state of conservation of these artefacts, often with damaged or missing areas, makes it difficult to perceive their original aesthetic value. To overcome these limitations, we propose an application of augmented reality able to support the observer in two ways: first, the application completes the missing surface of the mosaic or pavement by integrating the existent surface with a virtual reconstruction; second, it enables the analysis of the geometric pattern of the mosaic/pavement by overlaying virtual lines and geometric figures in order to explicit its geometric arrangements. The result is achieved via a custom Android application able to recognize and track the mosaic figure pattern and extra marker board, obtaining in that way a coordinate system used to render in real-time the reconstruction of the mosaic. Such rendering is overlaid to the video stream of the real scene. The application runs on a standard smartphone embedded in a Google Cardboard-compatible viewer and therefore is extremely affordable. As a case study, in order to reconstruct its aspects and to analyse its geometric pattern, we chose the roman mosaic re-found in Savignano sul Panaro (near Modena, Italy) in 2011, after 115 years from its first discovery, which is preserved less than half of its original 4.5 x 6.9 m surface.

Keywords: Augmented Reality , Cultural heritage , Geometric pattern , Photogrammetry , Real-time visualization , Roman mosaic

This paper proposes the integration of photogrammetric reconstruction, 3D modelling and augmented reality application in order to achieve the complete visualization of a stone sculpture even if highly damaged or fragmentary. The first part of the research aims to the reconstruction of the original aspect of an incomplete sculpture, by using photogrammetry techniques based on standard resolution photos and free software in order to obtain a first model; then, we integrate this model with other 3D digital data (from other sculptures of the same period) or with 3D modelling based on historical sources and views from historians, aiming to achieve the original aspect of the sculpture. The second part of the research consists of the embedding of the obtained model in a custom application able to render in real-time the 3D reconstruction of the lion. Then, the rendering is overlaid to the video stream of the real scene and, as a result, a complete 3D digital model of the sculpture is achieved and could be visualized through a VR viewer. As a case study, we focus on a Roman stone sculpture of a male lion conserved in the Museo Estense of Modena (Italy), which lacks of its head and its four legs. The original aspect of the lion may be achieved by integrating the damaged sculpture with other photogrammetric reconstructions of lions sculptures of the same period and with 3D model based on historical sources. Finally, the lion is visualized through an augmented reality application which digitally overlays the reconstructed models on the original one.

Keywords: Augmented Reality , Cultural heritage , Photogrammetry , Real-time visualization , Virtual modeling

Meniscectomy significantly change the kinematics of the knee joint by reducing the contact area between femoral condyles and the tibial plateau, but the shift in the contact area has been poorly described. The aim of our investigation was to measure the shift of the tibiofemoral contact area occurring after meniscectomy. We used laser scans combined to surface texturing for measuring the 3D position and area of the femoral and tibial surfaces involved in the joint. In particular, natural condyles (porcine model) were analysed and the reverse engineering approach was used for the interpretation of the results from compression tests and local force measurements in conjunction with staining techniques. The results suggested that laser scans combined to surface texturing may be considered as a powerful tool to investigate the stained contours of the contact area. Beside the largely documented reduction of contact area and local pressure increase, a shift of the centroid of the contact area toward the intercondylar notch was measured after meniscectomy. As a consequence of the contact area shift and pressure increase, cartilage degeneration close to the intercondylar notch may occur.

Keywords: Biomechanics , Centroid , Image analysis , Laser scanning , Surface texturing , Tibiofemoral contact area

In the Industry 4.0 and digital revolution era, the world of manufacturing industry is experiencing an innovative reconfiguration of design tools and methodologies, with a different approach to the production processes organization. The design philosophy is changing, integrating to engineering contribution interpretative aspects (design thinking), executive practices (design doing) and cognitive aspects (design cultures). The design becomes human-centered. The new Virtual Reality technologies allow to validate performances of designed products and production processes by means of virtual prototypes in a virtual simulated environment. This approach generates several benefits to the companies, in terms of costs and time, and allows optimizing the assembly line design and related workplaces, by improving workers' benefits too. This paper proposes an innovative method to validate the design of workplaces on automotive assembly lines in a virtual environment, based on ergonomic approach, according to ERGO - Uas system, applied by FCA (Fiat Chrysler Automobiles) groups, that integrates UAS method for measurement and EAWS method for biomechanical effort evaluation. Creating 3D virtual scenarios allows to carry on assembly tasks by virtual manikins in order to be evaluated from different points of view. In particular, data coming from the simulation can be used to assess several ergonomic indexes, improving safety, quality and design. The analysis is supported by the use of a motion capture system, developed by the University of Campania and composed of wearable inertial sensors, that estimates the attitude of fundamental human segments, using sensor fusion algorithms based on Kalman filtering. In this way, it is possible to make a further design validation, assessing the EAWS index basing on posture angles trends evaluated. This method can represent an innovation for human-centered design of workplace in developing new products, reducing costs and improving job quality.

Keywords: design , ergonomics , manufacturing , motion capture , product feasibility , simulation , Virtual reality

Methodologies for 3D face recognition which work in the presence of occlusions are core for the current needs in the field of identification of suspects, as criminals try to take advantage of the weaknesses among the implemented security systems by camouflaging themselves and occluding their face with eyeglasses, hair, hands, or covering their face with scarves and hats. Recent occlusion detection and restoration strategies for recognition purposes of 3D partially occluded faces with unforeseen objects are here presented in a literature review. The research community has worked on face recognition systems under controlled environments, but uncontrolled conditions have been investigated in a lesser extent. The paper details the experiments and databases used to handle the problem of occlusion and the results obtained by different authors. Lastly, a comparison of various techniques is presented and some conclusions are drawn referring to the best outcomes.

Keywords: 3D face analysis , Face detection , Face recognition , Facial occlusions , Facial restoration

This paper presents an approach for monitoring exercises of hand rehabilitation for past stroke patients. The developed solution uses a Leap Motion controller as hand-tracking device and embeds a supervised Machine Learning methodology. Support Vector Machine (SVM) is used in order to assess the correctness of a set of simple rehabilitation exercises performed with a single hand. The basic SVM model was extended with particular interest for defining feature vectors in a continues environment. The proposed method incorporated leap motion data, normalization of angles and gesture recognition. A software system was developed to provide patients with a set of exercise corrections and guidance for rehabilitation.

Keywords: Hand Rehabilitation , Leap Motion Controller , Stroke Patients , Support Vector Machines

Underwater images usually suffer from poor visibility, lack of contrast and colour casting, mainly due to light absorption and scattering. In literature, there are many algorithms aimed to enhance the quality of underwater images through different approaches. Our purpose was to identify an algorithm that performs well in different environmental conditions. We have selected some algorithms from the state of the art and we have employed them to enhance a dataset of images produced in various underwater sites, representing different environmental and illumination conditions. These enhanced images have been evaluated through some quantitative metrics. By analysing the results of these metrics, we tried to understand which of the selected algorithms performed better than the others. Another purpose of our research was to establish if a quantitative metric was enough to judge the behaviour of an underwater image enhancement algorithm. We aim to demonstrate that, even if the metrics can provide an indicative estimation of image quality, they could lead to inconsistent or erroneous evaluations.

Keywords: Automatic colour , Colour correction , Dehazing , Underwater image enhancement

Underwater sites are a challenging environment for augmented reality. Images taken under water are degraded in several ways, most importantly they are affected by unbalanced colors due to uneven absorption of light in each color channel, blurring and desaturation caused by turbidity, or noise due to the presence of larger organisms, plants, or bubbles. In this paper, we introduce a new method based on white balancing that enhances underwater images to improve the results of detection of markers. We compare this method with several image enhancement methods, and evaluate their performance when applied to the problem of detecting markers under water. Our results show that our method improves the detection in underwater environments while keeping the computation time low.

Objectives To investigate the influence of specific resin-composite, glass ceramic and glass ionomer cement (GIC) material combinations in a “multi-layer” technique to replace enamel and dentin in class II mesio-occlusal-distal (MOD) dental restorations using 3D-Finite Element Analysis (FEA). Methods Four 3D-FE models (A–D) of teeth, adhesively restored with different filling materials, were created and analyzed in comparison with a 3D model (E) of a sound lower molar. Models A, B & C had “multilayer” constructions, consisting of three layers: adhesive, dentin replacement and enamel replacement. Model A: had a low modulus (8 GPa) composite replacing dentin and a higher modulus (12 GPa) composite replacing enamel. Model B: had a GI cement replacing dentin and a higher modulus (12 GPa) composite replacing enamel. Model C: had a low modulus (8 GPa) composite replacing dentin and a very high modulus (70 GPa) inlay replacing enamel. Model D: had a lithium disilicate inlay replacing both dentin and enamel with a luting cement base-layer. Polymerization shrinkage effects were simulated and a load of 600 N was applied. All the materials were assumed to behave elastically throughout the entire deformation. Results Model A showed the highest stress distribution along all the adhesive interfaces of the shrinking resin-based materials with a critical condition and failure risk marginally and internally. Model D, by contrast, showed a more favorable performance than either of the multilayer groups (A–C). Stress and displacement plots showed an elastic response similar to that obtained for the sound tooth model. Model B and Model C performed according to their bilayer material properties. The use of a non-shrink dentin component simulating a GIC clearly affected the shrinkage stress at the basis of the Model B; while the bulk resin composite having a 12 GPa Young's modulus and linear polymerization shrinkage of 1% strongly influenced the biomechanical response in the bucco-lingual direction. Significance Direct resin-based composite materials applied in multilayer techniques to large class II cavities, with or without shrinking dentin layers, produced adverse FEA stress distributions and displacements. An indirect lithium disilicate inlay used to replace lost dentin and enamel in posterior restored teeth generated lower stress levels, within the limits of the elastic FEA model.

Keywords: CAD , Class II restorations , Finite element analysis , Image analysis , Materials properties

Objective To assess the effect of a ferrule design with specific post material-shape combinations on the mechanical behavior of post-restored canine teeth. Methods Micro-CT scan images of an intact canine were used to create a 3-D tessellated CAD model, from which the shapes of dentin, pulp and enamel were obtained and geometric models of post-endodontically restored teeth were created. Two types of 15 mm post were evaluated: a quartz fiber post with conical–tapered shape, and a carbon (C) fiber post with conical–cylindrical shape. The abutment was created around the coronal portion of the posts and 0.1 mm cement was added between prepared crown and abutment. Cement was also added between the post and root canal and a 0.25 mm periodontal ligament was modeled around the root. Four models were analysed by Finite Element (FE) Analysis: with/without a ferrule for both types of post material and shape. A load of 50 N was applied at 45° to the longitudinal axis of the tooth, acting on the palatal surface of the crown. The maximum normal stress criterion was adopted as a measure of potential damage. Results Models without a ferrule showed greater stresses (16.3 MPa) than those for models with a ferrule (9.2 MPa). With a ferrule, stress was uniformly distributed along the abutment and the root, with no critical stress concentration. In all models, the highest stresses were in the palatal wall of the root. Models with the C-fiber post had higher stress than models with the quartz fiber posts. The most uniform stress distribution was with the combination of ferrule and quartz fiber post. Significance The FE analysis confirmed a beneficial ferrule effect with the combination of ferrule and quartz fiber post, with tapered shape, affording no critical stress concentrations within the restored system.

Keywords: CAD , Dental materials , Endodontic treatment , Finite element analysis , Image analysis , Materials properties

The use of polymer composites has been increasing over the years and nowadays the requirements for designing high performance and lightweight fabrics and laminates for sail manufacturing have become more stringent than ever. The present paper offers an effective methodology that enhances the understanding of the influence of fibres orientation and arrangement of panels on sail performance. Constitutive characteristics of the ten commonly used sail cloths are experimentally measured and their influence on sail dynamic performance is compared using an aerodynamic approach. As expected also in industry 4.0 the method allows to control the production process and final product optimization.

Keywords: Aerodynamic coefficient , Apparent wind angle (AWA) , Apparent Wind Speed (AWS) , CFD analysis , Digital photogrammetry , RE , Turbulence model

Sail manufacture has undergone significant development due to sailing races like the America's Cup and the Volvo around the World Race. These competitions require advanced technologies to help increase sail performance. Hull design is fundamentally important but the sails (the only propulsion instrument) play a key role in dynamic of sailboats. Under aerodynamic loads, sail cloth deforms, the aerodynamic interaction is modified and the pressure on the sails is variously distributed resulting in performance inconsistencies. The interaction between fluid and structure necessitates a solution which combines aerodynamic and structural numerical simulations. Furthermore, in numerical simulations the aeroelastic sail characteristics must be known accurately. In this paper, the dynamic performance of a Spinnaker was studied. Digital photogrammetry was used to acquire the images, make the 3D reconstruction of the sail and validate the models in Computational Fluid Dynamics (CFD) analysis. Orthotropic constitutive characteristics of ten different sail cloths were measured by experimental test. The methodology allowed to compare dynamic performance in terms of forces, pressure and vibration for the different sail cloths and different fiber orientations.

Keywords: CFD analysis , Detached eddy simulations , Pressure distributions , Sail aerodynamics , Turbulence models

Definition of size, shape and location of defects into a mechanical component is of extreme importance in the manufacturing industry in general and particularly in high-tech applications, and in applications that can become dangerous due to the structural failure of mechanical components. In this paper, a laser-UT system has been used to define position and shape of internal defects in aluminum plates. An infrared pulsed laser is used to generate ultrasonic waves in a point of the plate and a CW laser interferometer is used as receiver to acquire the out-of plane displacements due to the ultrasonic waves in another point of the plate. The method consists of acquiring a B-Scan map on which some information on the defects in the mechanical component are visible. Storing the characteristics of the wave reflected by the defect and acquired in the B-Scan, the detection and the drawing of the defect is possible. The acquisition of the times of arrival of the waves reflected by the defect from the B-scan allows defining large parts of the shape of the defect. The times of arrival are acquired from the B-scan by analyzing the colour variations due to the wave reflected by the defect. The experiments operated from both sides of the plate allow drawing the defect in a virtual image of the plate section, from which the definition of defect shape and position can be determined.

Keywords: B-scan image analysis , Defect definition , Laser ultrasonic testing , NDE

The way a person moves, either in a plain walk or performing a specific task, tells a huge quantity of information about his/her physical and, eventually, neurologic condition. A large part of a physiotherapist work of assessment is based on the qualitative evaluation, mainly visual, of a person's movements, in terms of balance, speed, control, force and other parameters. This research work aims at providing personnel involved in the rehab process with a quantitative method to assess the way movements are performed. A numerical measure of the performance, actually, allows easier and more precise assessment, eliminating bias due to subjectivity. To accomplish this goal two steps are required: 3D acquisition of the movement using a Motion Capture (Mocap) system, and analysis of collected data to extract or elaborate the final outcome in the form requested by the medical staff. The paper shows the way Mocap acquisition are performed and data are analyzed in the application with people having a complete spinal cord injury and using a wheelchair. The method has been tested with eight volunteers in the rehabilitation department of the Hospital Papa Giovanni XXIII in Bergamo, Italy.

The Italian fashion industry is nowadays subject to radical transformation; therefore, it needs to remain competitive and, at the same time, innovate itself, in order to strengthen its position in the global market. An important opportunity of innovation can be the introduction of ICT technologies in the garment design process, which today is based on traditional methods and tools. Moreover, this innovation could be particularly important for online sales, in order to reduce the customers' doubts during purchasing. The research presented in this paper describes a framework for designing clothes as realistic 3D digital models and for allowing customers to evaluate the designed clothes by using realistic virtual mannequins of their bodies instead of the standard ones. A case study will be presented in the paper. The obtained results show that the framework can innovate the traditional garment design process and it could have a huge impact on fashion industry and customers behaviours.

Keywords: Body scanning , Cloth simulation , Design process , Motion capture , Virtual prototype

OBJECTIVE: The differential diagnosis of Parkinson's disease (PD) and other conditions, such as essential tremor and drug-induced parkinsonian syndrome or normal aging brain, represents a diagnostic challenge. 123I-FP-CIT brain SPET is able to contribute to the differential diagnosis. Semiquantitative analysis of radiopharmaceutical uptake in basal ganglia (caudate nuclei and putamina) is very useful to support the diagnostic process. An artificial neural network classifier using 123I-FP-CIT brain SPET data, a classification tree (CIT), was applied. CIT is an automatic classifier composed of a set of logical rules, organized as a decision tree to produce an optimised threshold based classification of data to provide discriminative cut-off values. We applied a CIT to 123I-FP-CIT brain SPET semiquantitave data, to obtain cut-off values of radiopharmaceutical uptake ratios in caudate nuclei and putamina with the aim to diagnose PD versus other conditions. SUBJECTS AND METHOD: We retrospectively investigated 187 patients undergoing 123I-FP-CIT brain SPET (Millenium VG, G.E.M.S.) with semiquantitative analysis performed with Basal Ganglia (BasGan) V2 software according to EANM guidelines; among them 113 resulted affected by PD (PD group) and 74 (N group) by other non parkinsonian conditions, such as Essential Tremor and drug-induced PD. PD group included 113 subjects (60M and 53F of age: 60-81yrs) having Hoehn and Yahr score (HY): 0.5-1.5; Unified Parkinson Disease Rating Scale (UPDRS) score: 6-38; N group included 74 subjects (36M and 38 F range of age 60-80 yrs). All subjects were clinically followed for at least 6-18 months to confirm the diagnosis. To examinate data obtained by using CIT, for each of the 1,000 experiments carried out, 10% of patients were randomly selected as the CIT training set, while the remaining 90% validated the trained CIT, and the percentage of the validation data correctly classified in the two groups of patients was computed. The expected performance of an "average performance CIT" was evaluated. RESULTS: For CIT, the probability of correct classification in patients with PD was 84.19±11.67% (mean±SD) and in N patients 93.48±6.95%. For CIT, the first decision rule provided a value for the right putamen of 2.32±0.16. This means that patients with right putamen values <2.32 were classified as having PD. Patients with putamen values ≥2.32 underwent further analysis. They were classified as N if the right putamen uptake value was ≥3.02 or if the value for the right putamen was <3.02 and the age was ≥67.5 years. Otherwise the patients were classified as having PD. Other similar rules on the values of both caudate nuclei and left putamen could be used to refine the classification, but in our data analysis of these data did not significantly contribute to the differential diagnosis. This could be due to an increased number of more severe patients with initial prevalence of left clinical symptoms having a worsening in right putamen uptake distribution. CONCLUSION: These results show that CIT was able to accurately classify PD and non-PD patients by means of 123I-FP-CIT brain SPET data and provided also cut-off values able to differentially diagnose these groups of patients. Right putamen uptake values resulted as the most discriminant to correctly classify our patients, probably due to a certain number of subjects with initial prevalence of left clinical symptoms. Finally, the selective evaluation of the group of subjects having putamen values ≥2.32 disclosed that age was a further important feature to classify patients for certain right putamen values.

Background/purpose: We investigate the use of skin texture features from the inner forearm as a means for personal identification. The forearm offers a number of potential advantages in that it is a fairly accessible area, and, compared with other zones such as fingertips, is less exposed to the elements and more shielded from wear. Methods: We extract and combine skin textural features from two imaging devices (optical and capacitive) with the aim of discriminating between different individuals. Skin texture images from 43 subjects were acquired from three different body parts (back of the hand, forearm and palm); testing used the two sensors either separately or in combination. Results: Skin texture features from the forearm proved effective for discriminating between different individuals with overall recognition accuracy approaching 96%. Conclusions: We found that skin texture features from the forearm are highly individual-specific and therefore suitable for personal identification. Interestingly, forearm skin texture features yielded significantly better accuracy compared to the skin of the back of the hand and of the palm of the same subjects.

Keywords: image processing , personal identification , skin texture , texture analysis