Image Processing & Applications

This paper describes the final assessment of a student with SLD who uses compensatory tools to verify the effectiveness of the strategies adopted to support the construction of an inclusive assessment environment. The proposed case study is part of long-term research conducted by the University of Macerata. New technologies for analyzing emotional feedback based on the analysis of facial expressions collected from both students and professors during the exam supported the observation. Results show how the application of the emotional recognition system allows researchers to reflect on the evaluation process, on the teacher's perception about compensatory tools and, more generally, on accommodation exams strategies for students with DSA.

Keywords: Emotional feedback , Evaluation strategies , Inclusive university teaching , University students Dyslexia

This paper introduces a new recommendation system for museums able to profile the visitors and propose them the most suitable exhibition path accordingly, to improve visitors' satisfaction. It consists of an interactive touch screen totem, which implements a USB camera and exploits Convolutional Neural Network to perform facial coding to measure visitors' emotions and estimate their age and gender. Based on the detected level of emotional valence, the system associates visitors with a profile and suggests them to visit a selection of five works of art, following a specific itinerary. An extensive experimentation lasting 2 months has been carried out at the Modern Art Museum “Palazzo Buonaccorsi” of Macerata. Results evidence that the proposed system can create an interactive and emotional link with the visitors, influencing their mood in the Pre-Experience phase and in the subsequent Post-Experience phase. In particular, they highlight that the proposed system, which aims at acting as emotional leverage, has been able to improve the positiveness of the emotions experienced by the visitors.

Keywords: Affective computing , Cultural heritage , Emotion recognition , Facial expression recognition

A time-domain filter is presented in this paper for the noise reduction of the camera images, then elaborated for the full-field 3D-DIC vibration analysis of mechanical components. The basic idea behind this filtering is to initially decompose the light signal of each pixel of the raw images, both for the left and the right cameras, with the Fast Fourier Transform (FFT). The signal is then reconstructed with the Inverse FFT by keeping the zero-order harmonic and the first harmonic and just discarding all the others, which only introduce noise contributions. The down-sampling approach is also used in combination with this filtering to reach high loading frequencies, even without (expensive) high-speed cameras, provided that a quite short exposure time is available. A cantilever aluminum plate was tested at 591 Hz, and a similar experiment was repeated on a bladed disk (or blisk) excited at high frequency, 6458 Hz, obtaining clear and smooth displacement maps even in the range of 10 microns. In the blisk application, a single blade with a significantly larger displacement amplitude was clearly observed, which could be interpreted as mistuning evidence.

Vibration measurements of turbomachinery components are of utmost importance to char-acterize the dynamic behavior of rotating machines, thus preventing undesired operating conditions. Local techniques such as strain gauges or laser Doppler vibrometers are usually adopted to collect vibration data. However, these approaches provide single-point and generally 1D measurements. The present work proposes an optical technique, which uses two low-speed cameras, a multimedia projector, and three-dimensional digital image correlation (3D-DIC) to provide full-field measurements of a bladed disk undergoing harmonic response analysis (i.e., pure sinusoidal excitation) in the kHz range. The proposed approach exploits a downsampling strategy to overcome the limitations introduced by low-speed cameras. The developed experimental setup was used to measure the response of a bladed disk subjected to an excitation frequency above 6 kHz, providing a deep insight in the deformed shapes, in terms of amplitude and phase distributions, which could not be feasible with single-point sensors. Results demonstrated the system's effectiveness in measuring amplitudes of few microns, also evidencing blade mistuning effects. A deeper insight into the deformed shape analysis was provided by considering the phase maps on the entire blisk geometry, and phase variation lines were observed on the blades for high excitation frequency.

Keywords: Bladed disk vibration , Digital image correlation , Downsampling , Low-speed cameras , Mistuning

Color schemes play a crucial role in blending virtual objects with the real environment. Good color schemes improve user's perception, which is of crucial importance for augmented reality. In this paper, we propose a set of novel methods based on the color harmonization methodology to recolor augmented reality content according to the real background. Three different strategies are proposed—harmonic, disharmonic, and balance—that allow for satisfying different needs in different settings depending on the application field. The first approach aims to harmonize the colors of virtual objects to make them consistent with the colors of the real background and reach a more pleasing effect to a human eye. The second approach, instead, can be adopted to generate a set of disharmonious colors with respect to real ones to be associated with the augmented virtual content to improve its distinctiveness from the real background. The third approach balances these goals by achieving a compromise between harmony and good visibility among virtual and real objects. Furthermore, the proposed re-coloring method is applied to three different case studies by adopting the three strategies to meet three different objectives, which are specific for each case study. Several parameters are calculated for each test, such as the covered area, the color distribution, and the set of generated colors. Results confirm the great potential of the proposed approaches to improve the AR visualization in different scenarios.

Keywords: Augmented reality , Color harmonization , Image processing

In physical rehabilitation, motion capture solutions are well-known but not as widespread as they could be. The main limit to their diffusion is not related to cost or usability but to the fact that the data generated when tracking a person must be elaborated according to the specific context and aim. This paper proposes a solution including customized motion capture and data elaboration with the aim of supporting medical personnel in the assessment of spinal cord-injured (SCI) patients using a wheelchair. The configuration of the full-body motion capturing system is based on an asymmetric 3 Microsoft Kinect v2 sensor layout that provides a path of up to 6 m, which is required to properly track the wheelchair. Data elaboration is focused on the automatic recognition of the pushing cycles and on plotting any kinematic parameter that may be interesting in the assessment. Five movements have been considered to evaluate the wheelchair propulsion: the humeral elevation, the horizontal abduction of the humerus, the humeral rotation, the elbow flexion and the trunk extension along the sagittal plane. More than 60 volunteers with a spinal cord injury were enrolled for testing the solution. To evaluate the reliability of the data computed with SCI APPlication (APP) for the pushing cycle analysis, the patients were subdivided in four groups according to the level of the spinal cord injury (i.e., high paraplegia, low paraplegia, C7 tetraplegia and C6 tetraplegia). For each group, the average value and the standard deviation were computed and a comparison with similar acquisitions performed with a high-end solution is shown. The measurements computed by the SCI-APP show a good reliability for analyzing the movements of SCI patients' propulsion wheelchair.

Keywords: Automatic pushing analysis , Markerless motion capture , RGB-D sensors , SCI patients

Motion capture (Mocap) is applied to motor rehabilitation of patients recovering from a trauma, a surgery, or other impairing conditions. Some rehabilitation exercises are easily tracked with low-cost technologies and a simple Mocap setup, while some others are extremely hard to track because they imply small movements and require high accuracy. In these last cases, the obvious solution is to use high performing motion tracking systems, but these devices are generally too expensive in the rehabilitation context. The aim of this paper is to provide a Mocap solution suitable for any kind of exercise but still based on low-cost sensors. This result can be reached embedding some artificial intelligence (AI), in particular a convolutional neural network (CNN), to gather a better outcome from the optical acquisition. The paper provides a methodology including the way to perform patient's tracking and to elaborate the data from infra-red sensors and from the red, green, blue (RGB) cameras in order to create a user-friendly application for physiotherapists. The approach has been tested with a known complex case concerning the rehabilitation of shoulders. The proposed solution succeeded in detecting small movements and incorrect patient behavior, as for instance, a compensatory elevation of the scapula during the lateral abduction of the arm. The approach evaluated by medical personnel provided good results and encouraged its application in different kinds of rehabilitation practices as well as in different fields where low-cost Mocap could be introduced.

Keywords: Artificial intelligence , Engineering informatics , Human computer interfaces/Interactions

Mobile devices and laptops are the main ICT tools to exchange information among people in the world. All the applications are designed by following a specific interaction style based either touchscreen or mouse and keyboard, which can be performed only with detailed movements of hands and fingers. Traditional interaction becomes difficult for elderly who have diseases limiting the hand motor skills, such as arthritis and brain stroke. The use of simple air gestures can be adopted as alternative interaction style to interact with smartphones, tablets and laptops. The aim of this research work is the development of an application that allows text writing using air gestures for people with limited hand motor skills. The application embeds several computer vision algorithms and convolutional neural networks software modules to detect and drawn alphanumeric characters and recognizing them using both mobile devices and laptops. The preliminary results obtained show that the approach is robust, and it can easily detect the alphanumeric characters written with the movement of the wrist..

Keywords: Elderly , Hand Motor Skills , OpenPose , TesserAct-OCR.

Motion capture (Mocap) systems are considered more and more interesting for the assessment of rehabilitation processes. In fact, medical personnel are increasingly demanding for technologies (possibly low-cost) to quantitatively measure and assess patients' improvements during rehabilitation exercises. In this paper, we focus the attention on the assessment of rehabilitation process for injured shoulders. This is particularly challenging because the recognition and the measurement of compensatory movements are very difficult during visual assessment and the movements of a shoulder are complex and arduous to be captured. The proposed solution integrates a low-cost Mocap system with video processing techniques to allow a quantitative evaluation of abduction, which is one of the first post-surgery exercises required for shoulder rehabilitation. The procedure is based on a set of open-source software tools to measure abduction and evaluate the correctness of the movement by detecting and measuring compensatory movements according to the parameters commonly considered by the physicians. Finally, a preliminary results and future works are presented and discussed.

Keywords: Embedded Knowledge , Medical assessment , Microsoft Kinect v2 , Motion Capture , Post-surgery shoulder

Dynamic characterization of vibrating targets represents a critical issue for many industrial fields. In this paper, a stereo-camera system integrated with a Digital Image Correlation (DIC) algorithm is proposed with the aim at performing 3D full-field vibration measurements in the range of kHz. The system exploits two industrial low-speed cameras, and the Nyquist-Shannon frequency limitation is overcome by a down-sampling approach under the hypothesis that the vibration signal is characterized by a single known frequency component. Experimental results obtained from the measurement of vibrational responses of a cantilever plate excited at three high-frequency resonance values (1121 Hz, 2956 Hz and 4010 Hz) are provided. A comparison with numerical analyses evidences the effectiveness of the proposed approach.

Keywords: Digital Image Correlation , Low-frame-rate cameras , Stereo-camera setup , Vibration measurement

In recent decades, increasing attention is being paid to the multidisciplinary approach that allows the performance of both a preventive conservation and a more invasive restoration action. In this context, the present study aims to acquire information and data from field surveys undertaken in San Domenico Church, Southern Calabria, in order to provide a tool for the recording and the inventory of damage and decay phenomena, and assess their causes and scale. The subsequent calculation of damage indices also provided useful information in order to allow the prioritization of conservation and preservation responses.

Keywords: Built heritage , Calabria , Computer graphics , Damage diagnosis , Decay , Italy , Photogrammetry

The paper presents the MOLUX project which aims to design and develop a system capable of supporting the underwater operators by means of innovative technologies for the diagnosis, recovery, management, conservation, and valorization of the archaeological and naturalistic assets in the marine environment. This system allows to increase the efficiency and the safety of the operations carried out in the underwater environment thanks to its capabilities: to make the diver aware of his/her position; to acquire geolocated data (images, videos, notes, environmental parameters); to optimize and monitor the planned dive route.

Keywords: Autonomous surface vehicle , Diver operation support , Monitoring , Underwater localization , Underwater tablet

Automatic painting classification by author, artistic genre and/or other attributes has generated considerable research interest in recent years. Being one of the visual features that mark the difference between artists and artistic genres, colour plays a fundamental role in this process. Colour is the result of the interaction among the intrinsic properties of the material, the illumination conditions and the response of the imaging device. Consequently, the same painting/artwork will look significantly different when imaged under varied conditions, which can be a potential source of bias for automated recognition procedures. One can compensate for such variations either via colour calibration or colour pre-processing. In this work we investigate the latter, and, in particular, evaluate the effectiveness of colour constancy and colour augmentation when coupled with hand-crafted and deep learning features for painting classification by artistic genre. In our experiments neither approach showed a clear advantage compared with no pre-processing at all. Colour constancy brought some improvement in certain cases, whereas colour augmentation virtually provided no benefit despite its adding a significant computational overload to the procedure.

Histological evaluation plays a major role in cancer diagnosis and treatment. The appearance of H&E-stained images can vary significantly as a consequence of differences in several factors, such as reagents, staining conditions, preparation procedure and image acquisition system. Such potential sources of noise can all have negative effects on computer-assisted classification. To minimize such artefacts and their potentially negative effects several color pre-processing methods have been proposed in the literature—for instance, color augmentation, color constancy, color deconvolution and color transfer. Still, little work has been done to investigate the efficacy of these methods on a quantitative basis. In this paper, we evaluated the effects of color constancy, deconvolution and transfer on automated classification of H&E-stained images representing different types of cancers—specifically breast, prostate, colorectal cancer and malignant lymphoma. Our results indicate that in most cases color pre-processing does not improve the classification accuracy, especially when coupled with color-based image descriptors. Some pre-processing methods, however, can be beneficial when used with some texture-based methods like Gabor filters and Local Binary Patterns.

Keywords: Color , H&E staining , Histology images , Texture

In this paper, we investigate the role of shape and texture features from 18F-FDG PET/CT to discriminate between benign and malignant solitary pulmonary nodules. To this end, we retrospectively evaluated cross-sectional data from 111 patients (64 males, 47 females, age = 67.5 ± 11.0) all with histologically confirmed benign (n = 39) or malignant (n = 72) solitary pulmonary nodules. Eighteen three-dimensional imaging features, including conventional, texture, and shape features from PET and CT were tested for significant differences (Wilcoxon-Mann-Withney) between the benign and malignant groups. Prediction models based on different feature sets and three classification strategies (Classification Tree, k-Nearest Neighbours, and Naïve Bayes) were also evaluated to assess the potential benefit of shape and texture features compared with conventional imaging features alone. Eight features from CT and 15 from PET were significantly different between the benign and malignant groups. Adding shape and texture features increased the performance of both the CT-based and PET-based prediction models with overall accuracy gain being 3.4–11.2 pp and 2.2–10.2 pp, respectively. In conclusion, we found that shape and texture features from 18F-FDG PET/CT can lead to a better discrimination between benign and malignant lung nodules by increasing the accuracy of the prediction models by an appreciable margin.

Keywords: Radiomics , Shape , Solitary pulmonary nodule , Texture

Transparent elements of buildings account for most of the variation in the spectrum of transmitted daylight. This behavior has a considerable impact on the luminous quality of the indoor environment, because the light spectral composition affects the perceived color and brightness of the illuminated objects. As a consequence, the effect of light colour and intensity on the behavior and health of the occupants is very important. Among the innovative transparent solutions for buildings, monolithic aerogel is one of the most promising, both in terms of lighting and thermal performance. In this context, the present study investigates the colour rendering properties of monolithic aerogel. For the analysis we employed a machine vision system based on a dome-shape illuminator, a digital camera, a chroma meter, and a colour checker. We evaluated the effect of the aerogel on colour rendering by measuring the RGB and CIE Lab coordinates of the reference colours of the colour checker with and without the aerogel. The results show that, in general, the presence of aerogel tends to shift all the colours towards a blue hue, the maximum variations occurring with dark tones of red, yellow, and green. Finally the Colour Rendering Index Ra was calculated for the examined monolithic aerogel by means of a methodology developed in accordance to the EN ISO 410 standard and CIE standard procedures. The obtained values prove a rather good visual quality of the monolithic aerogel (Ra = 85).

Keywords: CIE Lab coordinates , Colour rendering , Colour shift , Monolithic aerogel , RGB coordinates

Quantitative extraction of imaging features from medical scans ('radiomics') has attracted a lot of research attention in the last few years. The literature has consistently emphasized the potential use of radiomics for computer-assisted diagnosis, as well as for predicting survival and response to treatment. Radiomics is appealing in that it enables full-field analysis of the lesion, provides nearly real-time results, and is non-invasive. Still, a lot of studies suffer from a series of drawbacks such as lack of standardization and repeatability. Such limitations, along with the unmet demand for large enough image datasets for training the algorithms, are major hurdles that still limit the application of radiomics on a large scale. In this paper, we review the current developments, potential applications, limitations, and perspectives of PET/CT radiomics with specific focus on the management of patients with lung cancer.

Keywords: CT , Lung cancer , PET , Radiomics

The identification of tissue regions within histopathological images represents a fundamental step for diagnosis, patient stratification and follow-up. However, the huge amount of image data made available by the ever improving whole-slide imaging devices gives rise to a bottleneck in manual, microscopy-based evaluation. Furthermore, manual procedures generally show a significant intra- and/or inter-observer variability. In this scenario the objective of this chapter is to investigate the effectiveness of image features from last-generation, pre-trained convolutional networks against variants of Local Binary Patterns for classifying tissue sub-regions into meaningful classes such as epithelium, stroma, lymphocytes and necrosis. Experimenting with seven datasets of histopathological images we show that both classes of methods can be quite effective for the task, but with a noticeable superiority of descriptors based on convolutional neural networks. In particular, we show that these can be seamlessly integrated with standard classifiers (e.g. Support Vector Machines) to attain overall discrimination accuracy between 95 and 99%.

There is increasing evidence that shape and texture descriptors from imaging data could be used as image biomarkers for computer-assisted diagnosis and prognostication in a number of clinical conditions. It is believed that such quantitative features may help uncover patterns that would otherwise go unnoticed to the human eye, this way offering significant advantages against traditional visual interpretation. The objective of this paper is to provide an overview of the steps involved in the process – from image acquisition to feature extraction and classification. A significant part of the work deals with the description of the most common texture and shape features used in the literature; overall issues, perspectives and directions for future research are also discussed.

Keywords: Computer-assisted medicine , Radiomics , Shape , Texture

Partial orders are the natural mathematical structure for comparing multivariate data that, like colours, lack a natural order. We introduce a novel, general approach to defining rank features in colour spaces based on partial orders, and show that it is possible to generalise existing rank based descriptors by replacing the order relation over intensity values by suitable partial orders in colour space. In particular, we extend a classical descriptor (the Texture Spectrum) to work with partial orders. The effectiveness of the generalised descriptor is demonstrated through a set of image classification experiments on 10 datasets of colour texture images. The results show that the partial-order version in colour space outperforms the grey-scale classic descriptor while maintaining the same number of features.

Keywords: Hand-designed image descriptors , Mathematics of colour and texture , Partial orders , Rank features