This paper presents an implementation of RoSA, a Robot System Assistant, for safe and intuitive human-machine interaction. The interaction modalities were chosen and previously reviewed using a Wizard of Oz study emphasizing a strong propensity for speech and pointing gestures. Based on these findings, we design and implement a new multi-modal system for contactless human-machine interaction based on speech, facial, and gesture recognition. We evaluate our proposed system in an extensive study with multiple subjects to examine the user experience and interaction efficiency. It reports that our method achieves similar usability scores compared to the entirely human remote-controlled robot interaction in our Wizard of Oz study. Furthermore, our framework’s implementation is based on the Robot Operating System (ROS), allowing modularity and extendability for our multi-device and multi-user method.

The ability to measure the engagement level of humans interacting with robots paves the way towards intuitive and safe human-robot interaction. Recent approaches achieve reasonable progress in predicting human engagement in physically situated environments. However, engagement estimation is still a challenging problem especially in an open-world environment due to the difficulty of creating and monitoring a variety of human social cues in real-time. Furthermore, the interactions may involve a group of subjects interacting simultaneously with the robot, which increases the prediction complexity. In this paper, we design a real-time engagement estimation system for humans interacting with robots with generalization capability. We propose to estimate engagement using a three-stage approach based on a combination of learning-based and rule-based approaches. Firstly, state-of-the-art deep learning methods are used to extract engagement features from input frames. Then, a simple neural network is used to estimate the focus of attention score by incorporating gaze and head pose features and assigning this score to all subjects in the scene using a face recognition algorithm. Finally, a rule-based classification approach is used to predict the engagement state of the subject to initiate/terminate the interaction with the robot. To effectively evaluate our system, we access our approach for each phase separately. Additionally, we use an online evaluation study in which subjects are allowed to interact freely with an industrial robot. Our model achieves an average of 96%, 90%, and 93% precision, recall, and F-score respectively.

Recently, face recognition became a key element in social cognition which is used in various applications including human–robot interaction (HRI), pedestrian identification, and surveillance systems. Deep convolutional neural networks (CNNs) have achieved notable progress in recognizing faces. However, achieving accurate and real-time face recognition is still a challenging problem, especially in unconstrained environments due to occlusion, lighting conditions, and the diversity in head poses. In this paper, we present a robust face recognition and tracking framework in unconstrained settings. We developed our framework based on lightweight CNNs for all face recognition stages, including face detection, alignment and feature extraction, to achieve higher accuracies in these challenging circumstances while maintaining the real-time capabilities required for HRI systems. To maintain the accuracy, a single-shot multi-level face localization in the wild (RetinaFace) is utilized for face detection, and additive angular margin loss (ArcFace) is employed for recognition. For further enhancement, we introduce a face tracking algorithm that combines the information from tracked faces with the recognized identity to use in the further frames. This tracking algorithm improves the overall processing time and accuracy. The proposed system performance is tested in real-time experiments applied in an HRI study. Our proposed framework achieves real-time capabilities with an average of 99%, 95%, and 97% precision, recall, and F-score respectively. In addition, we implemented our system as a modular ROS package that makes it straightforward for integration in different real-world HRI systems.

The rapid advancement of deep Convolutional Neural Networks (CNNs) has led to remarkable progress in computer vision, contributing to the development of numerous face verification architectures. However, the inherent complexity of these architectures, often characterized by millions of parameters and substantial computational demands, presents significant challenges for deployment on resource-constrained devices. To address these challenges, we introduce RobFaceNet, a robust and efficient CNN designed explicitly for face recognition (FR). The proposed RobFaceNet optimizes accuracy while preserving computational efficiency, a balance achieved by incorporating multiple features and attention mechanisms. These features include both low-level and high-level attributes extracted from input face images and aggregated from multiple levels. Additionally, the model incorporates a newly developed bottleneck that integrates both channel and spatial attention mechanisms. The combination of multiple features and attention mechanisms enables the network to capture more significant facial features from the images, thereby enhancing its robustness and the quality of facial feature extraction. Experimental results across state-of-the-art FR datasets demonstrate that our RobFaceNet achieves higher recognition performance. For instance, RobFaceNet achieves 95.95% and 92.23% on the CA-LFW and CP-LFW datasets, respectively, compared to 95.45% and 92.08% for very deep ArcFace model. Meanwhile, RobFaceNet exhibits a more lightweight model complexity. In terms of computation cost, RobFaceNet has 337M Floating Point Operations Per Second (FLOPs) compared to ArcFace’s 24211M, with only 3% of the parameters. Consequently, RobFaceNet is well-suited for deployment across various platforms, including robots, embedded systems, and mobile devices.

Deep feature learning has become crucial in large-scale face recognition, and margin-based loss functions have demonstrated impressive success in this field. These methods aim to enhance the discriminative power of the softmax loss by increasing the feature margin between different classes. These methods assume class balance, where a fixed margin is sufficient to squeeze intra-class variation equally. However, real-face datasets often exhibit imbalanced classes, where the fixed margin is suboptimal, limiting the discriminative power and generalizability of the face recognition model. Furthermore, margin-based approaches typically focus on enhancing discrimination either in the angle or cosine space, emphasizing one boundary while disregarding the other. To overcome these limitations, we propose a joint adaptive margins loss function (JAMsFace) that learns class-related margins for both angular and cosine spaces. This approach allows adaptive margin penalties to adjust adaptively for different classes. We explain and analyze the proposed JAMsFace geometrically and present comprehensive experiments on multiple face recognition benchmarks. The results show that JAMsFace outperforms existing face recognition losses in mainstream face recognition tasks. Specifically, JAMsFace advances the state-of-the-art face recognition performance on LFW, CPLFW, and CFP-FP and achieves comparable results on CALFW and AgeDB-30. Furthermore, for the challenging IJB-B and IJB-C benchmarks, JAMsFace achieves impressive true acceptance rates (TARs) of 89.09% and 91.81% at a false acceptance rate (FAR) of 1e-4, respectively.

Face and person detection are important tasks in computer vision, as they represent the first component in many recognition systems, such as face recognition, facial expression analysis, body pose estimation, face attribute detection, or human action recognition. Thereby, their detection rate and runtime are crucial for the performance of the overall system. In this paper, we combine both face and person detection in one framework with the goal of reaching a detection performance that is competitive to the state of the art of lightweight object-specific networks while maintaining real-time processing speed for both detection tasks together. In order to combine face and person detection in one network, we applied multi-task learning. The difficulty lies in the fact that no datasets are available that contain both face as well as person annotations. Since we did not have the resources to manually annotate the datasets, as it is very time-consuming and automatic generation of ground truths results in annotations of poor quality, we solve this issue algorithmically by applying a special training procedure and network architecture without the need of creating new labels. Our newly developed method called Simultaneous Face and Person Detection (SFPD) is able to detect persons and faces with 40 frames per second. Because of this good trade-off between detection performance and inference time, SFPD represents a useful and valuable real-time framework especially for a multitude of real-world applications such as, e.g., human–robot interaction.