Filmmakers and editors have empirically developed techniques to ensure the spatiotemporal continuity of a film's narration. In terms of time, editing techniques (e.g., elliptical, overlapping, or cut minimization) allow for the manipulation of the perceived duration of events as they unfold on screen. More specifically, a scene can be edited to be time compressed, expanded, or real-time in terms of its perceived duration. Despite the consistent application of these techniques in filmmaking, their perceptual outcomes have not been experimentally validated. Given that viewing a film is experienced as a precise simulation of the physical world, the use of cinematic material to examine aspects of time perception allows for experimentation with high ecological validity, while filmmakers gain more insight on how empirically developed techniques influence viewers' time percept. Here, we investigated how such time manipulation techniques of an action affect a scene's perceived duration. Specifically, we presented videos depicting different actions (e.g., a woman talking on the phone), edited according to the techniques applied for temporal manipulation and asked participants to make verbal estimations of the presented scenes' perceived durations. Analysis of data revealed that the duration of expanded scenes was significantly overestimated as compared to that of compressed and real-time scenes, as was the duration of real-time scenes as compared to that of compressed scenes. Therefore, our results validate the empirical techniques applied for the modulation of a scene's perceived duration. We also found interactions on time estimates of scene type and editing technique as a function of the characteristics and the action of the scene presented. Thus, these findings add to the discussion that the content and characteristics of a scene, along with the editing technique applied, can also modulate perceived duration. Our findings are discussed by considering current timing frameworks, as well as attentional saliency algorithms measuring the visual saliency of the presented stimuli.

DensePose provides a pixel-accurate association of images with 3D mesh coordinates, but does not provide a 3D mesh, while Human Mesh Reconstruction (HMR) systems have high 2D reprojection error, as measured by DensePose localization metrics. In this work we introduce MeshPose to jointly tackle DensePose and HMR. For this we first introduce new losses that allow us to use weak DensePose supervision to accurately localize in 2D a subset of the mesh vertices ('VertexPose'). We then lift these vertices to 3D, yielding a low-poly body mesh ('MeshPose'). Our system is trained in an end-to-end manner and is the first HMR method to attain competitive DensePose accuracy, while also being lightweight and amenable to efficient inference, making it suitable for real-time AR applications.

We introduce ViDaS, a two-stream, fully convolutional Video, Depth-Aware Saliency network to address the problem of attention modeling ``in-the-wild", via saliency prediction in videos. Contrary to existing visual saliency approaches using only RGB frames as input, our network employs also depth as an additional modality. The network consists of two visual streams, one for the RGB frames, and one for the depth frames. Both streams follow an encoder-decoder approach and are fused to obtain a final saliency map. The network is trained end-to-end and is evaluated in a variety of different databases with eye-tracking data, containing a wide range of video content. Although the publicly available datasets do not contain depth, we estimate it using three different state-of-the-art methods, to enable comparisons and a deeper insight. Our method outperforms in most cases state-of-the-art models and our RGB-only variant, which indicates that depth can be beneficial to accurately estimating saliency in videos displayed on a 2D screen. Depth has been widely used to assist salient object detection problems, where it has been proven to be very beneficial. Our problem though differs significantly from salient object detection, since it is not restricted to specific salient objects, but predicts human attention in a more general aspect. These two problems not only have different objectives, but also different ground truth data and evaluation metrics. To our best knowledge, this is the first competitive deep learning video saliency estimation approach that combines both RGB and Depth features to address the general problem of saliency estimation ``in-the-wild". The code will be publicly released.

In this paper, we present an integrated robotic system capable of participating in and performing a wide range of educational and entertainment tasks collaborating with one or more children. The system, called ChildBot, features multimodal perception modules and multiple robotic agents that monitor the interaction environment and can robustly coordinate complex Child–Robot Interaction use-cases. In order to validate the effectiveness of the system and its integrated modules, we have conducted multiple experiments with a total of 52 children. Our results show improved perception capabilities in comparison to our earlier works that ChildBot was based on. In addition, we have conducted a preliminary user experience study, employing some educational/entertainment tasks, that yields encouraging results regarding the technical validity of our system and initial insights on the user experience with it.

Robots can play a significant role as assistive devices for people with movement impairment and mild cognitive deficit. In this paper we present an overview of the lightweight i-Walk intelligent robotic rollator, which offers cognitive and mobility assistance to the elderly and to people with light to moderate mobility impairment. The utility, usability, safety and technical performance of the device is investigated through a clinical study, which took place at a rehabilitation center in Greece involving real patients with mild to moderate cognitive and mobility impairment. This first evaluation study comprised a set of scenarios in a number of pre-defined use cases, including physical rehabilitation exercises, as well as mobility and ambulation involved in typical daily living activities of the patients. The design and implementation of this study is discussed in detail, along with the obtained results, which include both an objective and a subjective evaluation of the system operation, based on a set of technical performance measures and a validated questionnaire for the analysis of qualitative data, respectively. The study shows that the technical modules performed satisfactory under real conditions, and that the users generally hold very positive views of the platform, considering it safe and reliable.

Robotic rollators can play a significant role as assistive devices for people with impaired movement and mild cognitive deficit. This paper presents an overview of the i-Walk concept; an intelligent robotic rollator offering cognitive and ambulatory assistance to people with light to moderate movement impairment, such as the elderly. We discuss the two robotic prototypes being developed, their various novel functionalities, system architecture, modules and function scope, and present preliminary experimental results with actual users.

We present the i-Walk system, a novel framework for intelligent mobility assistance applications. The proposed system is capable of automatically understanding human activity, assessing mobility and rehabilitation progress, recognizing human intentions and communicating with the patients by giving meaningful feedback. To this end, multiple sensors, i.e. cameras, microphones, lasers, provide multimodal data in order to allow for user monitoring, while state-of-the-art and beyond algorithms have been developed and integrated into the system to enable recognition, interaction and assessment. More specifically, i-Walk performs in real-time and consists of four main sub-modules that interact automatically to provide speech understanding, activity recognition, mobility analysis and multimodal communication for seamless HRI. The i-Walk assessment system is evaluated on a database of healthy subjects and patients, who participated in carefully designed experimental scenarios that cover essential needs of rehabilitation. The presented results highlight the efficacy of the proposed framework to endow personal assistants with intelligence.

We present a novel multi-attentional convolutional architecture to tackle the problem of real-time RGB-D 6D object pose tracking of single, known objects. Such a problem poses multiple challenges originating both from the objects' nature and their interaction with their environment, which previous approaches have failed to fully address. The proposed framework encapsulates methods for background clutter and occlusion handling by integrating multiple parallel soft spatial attention modules into a multitask Convolutional Neural Network (CNN) architecture. Moreover, we consider the special geometrical properties of both the object's 3D model and the pose space, and we use a more sophisticated approach for data augmentation during training. The provided experimental results confirm the effectiveness of the proposed multi-attentional architecture, as it improves the State-of-the-Art (SoA) tracking performance by an average score of 34.03% for translation and 40.01% for rotation, when tested on the most complete dataset designed, up to date,for the problem of RGB-D object tracking.

Background: Gesture-based human-robot interaction (HRI) depends on the technical performance of the robot-integrated gesture recognition system (GRS) and on the gestural performance of the robot user, which has been shown to be rather low in older adults. Training of gestural commands (GCs) might improve the quality of older users' input for gesture-based HRI, which in turn may lead to an overall improved HRI. Objective: To evaluate the effects of a user training on gesture-based HRI between an assistive bathing robot and potential elderly robot users. Methods: Twenty-five older adults with bathing disability participated in this quasi-experimental, single-group, pre-/post-test study and underwent a specific user training (10-15 min) on GCs for HRI with the assistive bathing robot. Outcomes measured before and after training included participants' gestural performance assessed by a scoring method of an established test of gesture production (TULIA) and sensor-based gestural performance (SGP) scores derived from the GRS-recorded data, and robot's command recognition rate (CRR). Results: Gestural performance (TULIA = +57.1 +/- 56.2 %, SGP scores = +41.1 +/- 74.4 %) and CRR (+31.9 +/- 51.2 %) significantly improved over training (p < .001). Improvements in gestural performance and CRR were highly associated with each other (r = 0.80-0.81, p < .001). Participants with lower initial gestural performance and higher gerontechnology anxiety benefited most from the training. Conclusions: Our study highlights that training in gesture-based HRI with an assistive bathing robot is highly beneficial for the quality of older users' GCs, leading to higher CRRs of the robot-integrated GRS, and thus to an overall improved HRI.

Robotic rollators can play a significant role as assistive devices for people with impaired movement and mild cognitive deficit. This paper presents an overview of the i-Walk concept; an intelligent robotic rollator offering cognitive and ambulatory assistance to people with light to moderate movement impairment, such as the elderly. We discuss the two robotic prototypes being developed, their various novel functionalities, system architecture, modules and function scope, and present preliminary experimental results with actual users.

In this paper we present a prototype integrated robotic system, the I-Support bathing robot, that aims at supporting new aspects of assisted daily-living activities on a real-life scenario. The paper focuses on describing and evaluating key novel technological features of the system, with the emphasis on cognitive human–robot interaction modules and their evaluation through a series of clinical validation studies. The I-Support project on its whole has envisioned the development of an innovative, modular, ICT-supported service robotic system that assists frail seniors to safely and independently complete an entire sequence of physically and cognitively demanding bathing tasks, such as properly washing their back and their lower limbs. A variety of innovative technologies have been researched and a set of advanced modules of sensing, cognition, actuation and control have been developed and seamlessly integrated to enable the system to adapt to the target population abilities. These technologies include: human activity monitoring and recognition, adaptation of a motorized chair for safe transfer of the elderly in and out the bathing cabin, a context awareness system that provides full environmental awareness, as well as a prototype soft robotic arm and a set of user-adaptive robot motion planning and control algorithms. This paper focuses in particular on the multimodal action recognition system, developed to monitor, analyze and predict user actions with a high level of accuracy and detail in real-time, which are then interpreted as robotic tasks. In the same framework, the analysis of human actions that have become available through the project’s multimodal audio–gestural dataset, has led to the successful modeling of Human–Robot Communication, achieving an effective and natural interaction between users and the assistive robotic platform. In order to evaluate the I-Support system, two multinational validation studies were conducted under realistic operating conditions in two clinical pilot sites. Some of the findings of these studies are presented and analyzed in the paper, showing good results in terms of: (i) high acceptability regarding the system usability by this particularly challenging target group, the elderly end-users, and (ii) overall task effectiveness of the system in different operating modes.

We introduce STAViS, a spatio-temporal audiovisual saliency network that combines spatio-temporal visual and auditory information in order to efficiently address the problem of saliency estimation in videos. Our approach employs a single network that combines visual saliency and auditory features and learns to appropriately localize sound sources and to fuse the two saliencies in order to obtain a final saliency map. The network has been designed, trained end-to-end, and evaluated on six different databases that contain audiovisual eye-tracking data of a large variety of videos. We compare our method against 8 different state-of-the-art visual saliency models. Evaluation results across databases indicate that our STAViS model outperforms our visual only variant as well as the other state-of-the-art models in the majority of cases. Also, the consistently good performance it achieves for all databases indicates that it is appropriate for estimating saliency "in-the-wild". The code is available at https://github.com/atsiami/STAViS.

In this work, we present a novel framework for on-line human gait stability prediction of the elderly users of an intelligent robotic rollator using Long Short Term Memory (LSTM) networks, fusing multimodal RGB-D and Laser Range Finder (LRF) data from non-wearable sensors. A Deep Learning (DL) based approach is used for the upper body pose estimation. The detected pose is used for estimating the body Center of Mass (CoM) using Unscented Kalman Filter (UKF). An Augmented Gait State Estimation framework exploits the LRF data to estimate the legs' positions and the respective gait phase. These estimates are the inputs of an encoder-decoder sequence to sequence model which predicts the gait stability state as Safe or Fall Risk walking. It is validated with data from real patients, by exploring different network architectures, hyperparameter settings and by comparing the proposed method with other baselines. The presented LSTM-based human gait stability predictor is shown to provide robust predictions of the human stability state, and thus has the potential to be integrated into a general user-adaptive control architecture as a fall-risk alarm.

Detecting visual relationships, i.e. triplets, is a challenging Scene Understanding task approached in the past via linguistic priors or spatial information in a single feature branch. We introduce a new deeply supervised two-branch architecture, the Multimodal Attentional Translation Embeddings, where the visual features of each branch are driven by a multimodal attentional mechanism that exploits spatio-linguistic similarities in a low-dimensional space. We present a variety of experiments comparing against all related approaches in the literature, as well as by re-implementing and fine-tuning several of them. Results on the commonly employed VRD dataset [1] show that the proposed method clearly outperforms all others, while we also justify our claims both quantitatively and qualitatively.

In this letter, we address the problem of multi-cue affect recognition in challenging scenarios such as child–robot interaction. Toward this goal we propose a method for automatic recognition of affect that leverages body expressions alongside facial ones, as opposed to traditional methods that typically focus only on the latter. Our deep-learning based method uses hierarchical multi-label annotations and multi-stage losses, can be trained both jointly and separately, and offers us computational models for both individual modalities, as well as for the whole body emotion. We evaluate our method on a challenging child–robot interaction database of emotional expressions collected by us, as well as on the GEneva multimodal emotion portrayal public database of acted emotions by adults, and show that the proposed method achieves significantly better results than facial-only expression baselines.

In this work, we tackle the problem of child engagement estimation while children freely interact with a robot in a friendly, room-like environment. We propose a deep learning-based multi-view solution that takes advantage of recent developments in human pose detection. We extract the child's pose from different RGB-D cameras placed regularly in the room, fuse the results and feed them to a deep Neural Network (NN) trained for classifying engagement levels. The deep network contains a recurrent layer, in order to exploit the rich temporal information contained in the pose data. The resulting method outperforms a number of baseline classifiers and provides a promising tool for better automatic understanding of a child's attitude, interest and attention while cooperating with a robot. The goal is to integrate this model in nextgeneration social robots as an attention monitoring tool during various Child Robot Interaction (CRI) tasks both for Typically Developed (TD) children and children affected by autism (ASD).

In this work we propose a multi-task spatio-temporal network, called SUSiNet, that can jointly tackle the spatio-temporal problems of saliency estimation, action recognition and video summarization. Our approach employs a single network that is jointly end-to-end trained for all tasks with multiple and diverse datasets related to the exploring tasks. The proposed network uses a unified architecture that includes global and task specific layer and produces multiple output types, i.e., saliency maps or classification labels, by employing the same video input. Moreover, one additional contribution is that the proposed network can be deeply supervised through an attention module that is related to human attention as it is expressed by eye-tracking data. From the extensive evaluation, on seven different datasets, we have observed that the multi-task network performs as well as the state-of-the-art single-task methods (or in some cases better), while it requires less computational budget than having one independent network per each task.

Human attention is highly influenced by multi-modal combinations of perceived sensory information and especially audiovisual information. Although systematic behavioral experiments have provided evidence that human attention is multi-modal, most bottom-up computational attention models, namely saliency models for fixation prediction, focus on visual information, largely ignoring auditory input. In this work, we aim to bridge the gap between findings from neuroscience concerning audiovisual attention and the computational attention modeling, by creating a 2-D bottom-up audiovisual saliency model. We experiment with various fusion schemes for integrating state-of-the-art auditory and visual saliency models in a single audiovisual attention/saliency model based on behavioral findings, that we validate in two experimental levels: (1) using results from behavioral experiments aiming to reproduce the results in a mostly qualitative manner and to ensure that our modeling is in line with behavioral findings, and (2) using 6 different databases with audiovisual human eye-tracking data. For this last purpose, we have also collected eye-tracking data for two databases: ETMD, a movie database that contains highly edited videos (movie clips), and SumMe, a database that contains unstructured and unedited user videos. Experimental results indicate that our proposed audiovisual fusion schemes in most cases improve performance compared to visual-only models, without any prior knowledge of the video/audio content. Also, they can be generalized and applied to any auditory saliency model and any visual spatio-temporal saliency model.

We find that most Scene Graph Generation approaches suffer from two limitations as they: 1) use generic attention mechanisms and dataset-specific statistics that supersede visual features and 2) treat "no interaction" as an extra, both noisy and dominant, class and prune graph edges manually or applying simple filters. As a result, such approaches do not scale up on different settings and specifications. We propose a three-stage pipeline that employs Multi-Head Attention driven by language and spatial features, Translation Embeddings and Multi-Tasking to detect an interacting pair of objects. Our attentional scheme is able to maximize the visual features' interpretability, as well as to capture the nature of datasets of different scales, while multi-tasking robustly resolves the bias of the background class. We present an experimental overview of the related literature, unveil a multitude of evaluation inconsistencies and provide quantitative and qualitative support with experiments on a variety of datasets, where our approach performs on par or even outperforms current state-of-the-art.

The integration of visual perception to robotic systems is a key research area in recent years. Advances in modern computer vision techniques along with the development of faster and more accurate visual sensors led to the emergence of new methods for robotic visual perception [1]. One area of research is the development of robotic assistive vision for Human-Robot Interaction (HRI) systems [2], [3]. The rapid increase of people with special needs, such as the elderly population, and the simultaneous reduction of personal care staff, reinforce the need for robotic assistants [4], [5]. There are many challenges in this area including the familiarity of these users with new technologies and the domain specific datasets, which are required for training user oriented models. Nowadays, modern assistive and social human-robot interaction requires the multimodal communication with speech, gestures and human movements so as to enhance the classic interaction with only spoken commands.