Our comprehension of video streams depicting human activities is naturally multifaceted: in just a few moments, we can grasp what is happening, identify the relevance and interactions of objects in the scene, and forecast what will happen soon, everything all at once. To endow autonomous systems with such a holistic perception, learning how to correlate concepts, abstract knowledge across diverse tasks, and leverage tasks synergies when learning novel skills is essential. A significant step in this direction is EgoPack, a unified framework for understanding human activities across diverse tasks with minimal overhead. EgoPack promotes information sharing and collaboration among downstream tasks, essential for efficiently learning new skills. In this paper, we introduce Hier-EgoPack, which advances EgoPack by enabling reasoning also across diverse temporal granularities, which expands its applicability to a broader range of downstream tasks. To achieve this, we propose a novel hierarchical architecture for temporal reasoning equipped with a GNN layer specifically designed to tackle the challenges of multi-granularity reasoning effectively. We evaluate our approach on multiple Ego4D benchmarks involving both clip-level and frame-level reasoning, demonstrating how our hierarchical unified architecture effectively solves these diverse tasks simultaneously.

Referring Video Object Segmentation (RVOS) relies on natural language expressions to segment an object in a video clip. Existing methods restrict reasoning either to independent short clips, losing global context, or process the entire video offline, impairing their application in a streaming fashion. In this work, we aim to surpass these limitations and design an RVOS method capable of effectively operating in streaming-like scenarios while retaining contextual information from past frames. We build upon the Segment-Anything 2 (SAM2) model, that provides robust segmentation and tracking capabilities and is naturally suited for streaming processing. We make SAM2 wiser, by empowering it with natural language understanding and explicit temporal modeling at the feature extraction stage, without fine-tuning its weights, and without outsourcing modality interaction to external models. To this end, we introduce a novel adapter module that injects temporal information and multi-modal cues in the feature extraction process. We further reveal the phenomenon of tracking bias in SAM2 and propose a learnable module to adjust its tracking focus when the current frame features suggest a new object more aligned with the caption. Our proposed method, SAMWISE, achieves state-of-the-art across various benchmarks, by adding a negligible overhead of less than 5 M parameters. Code is available at https://github.com/ClaudiaCuttano/SAMWISE.

Vision Transformers (ViTs) have shown remarkable performance and scalability across various computer vision tasks. To apply single-scale ViTs to image segmentation, existing methods adopt a convolutional adapter to generate multi-scale features, a pixel decoder to fuse these features, and a Transformer decoder that uses the fused features to make predictions. In this paper, we show that the inductive biases introduced by these task-specific components can instead be learned by the ViT itself, given sufficiently large models and extensive pre-training. Based on these findings, we introduce the Encoder-Only Mask Transformer (EoMT), which repurposes the plain ViT architecture to conduct image segmentation. With large-scale models and pre-training, EoMT obtains a segmentation accuracy similar to state-of-the-art models that use task-specific components. At the same time, EoMT is significantly faster than these methods due to its architectural simplicity, e.g., up to 4 × faster with ViT-L. Across a range of model sizes, EoMT demonstrates an optimal balance between segmentation accuracy and prediction speed, suggesting that compute resources are better spent on scaling the ViT itself rather than adding architectural complexity. Code: https://www.tue-mps.org/eomt/.

Human activities are particularly complex and variable, and this makes challenging for deep learning models to reason about them. However, we note that such variability does have an underlying structure, composed of a hierarchy of patterns of related actions. We argue that such structure can emerge naturally from unscripted videos of human activities, and can be leveraged to better reason about their content. We present HiERO, a weakly-supervised method to enrich video segments features with the corresponding hierarchical activity threads. By aligning video clips with their narrated descriptions, HiERO infers contextual, semantic and temporal reasoning with an hierarchical architecture. We prove the potential of our enriched features with multiple video-text alignment benchmarks (EgoMCQ, EgoNLQ) with minimal additional training, and in zero-shot for procedure learning tasks (EgoProceL and Ego4D Goal-Step). Notably, HiERO achieves state-of-the-art performance in all the benchmarks, and for procedure learning tasks it outperforms fully-supervised methods by a large margin (+12.5% F1 on EgoProceL) in zero shot. Our results prove the relevance of using knowledge of the hierarchy of human activities for multiple reasoning tasks in egocentric vision.

Few-shot segmentation aims to segment unseen categories from just a handful of annotated examples. This requires mechanisms to identify semantically related objects across images and accurately produce masks. We note that Segment Anything 2 (SAM2), with its prompt-and-propagate mechanism, provides strong segmentation capabilities and a built-in feature matching process. However, we show that its representations are entangled with task-specific cues optimized for object tracking, which impairs its use for tasks requiring higher level semantic understanding. Our key insight is that, despite its class-agnostic pretraining, SAM2 already encodes rich semantic structure in its features. We propose SANSA (Semantically AligNed Segment Anything 2), a framework that makes this latent structure explicit, and repurposes SAM2 for few-shot segmentation through minimal task-specific modifications. SANSA achieves state-of-the-art on few-shot segmentation benchmarks designed to assess generalization and outperforms generalist methods in the popular in-context setting. Additionally, it supports flexible promptable interaction via points, boxes, or scribbles, and remains significantly faster and more compact than prior approaches. Code at: https://github.com/ClaudiaCuttano/SANSA.

Recognizing human activities from visual inputs, particularly through a first-person viewpoint, is essential for enabling robots to replicate human behavior. Egocentric vision, characterized by cameras worn by observers, captures diverse changes in illumination, viewpoint, and environment. This variability leads to a notable drop in the performance of Egocentric Action Recognition models when tested in environments not seen during training. In this paper, we tackle these challenges by proposing a domain generalization approach for Egocentric Action Recognition. Our insight is that action sequences often reflect consistent user intent across visual domains. By leveraging action sequences , we aim to enhance the model’s generalization ability across unseen environments. Our proposed method, named SeqDG, introduces a visual-text sequence reconstruction objective (SeqRec) that uses contextual cues from both text and visual inputs to reconstruct the central action of the sequence. Additionally, we enhance the model’s robustness by training it on mixed sequences of actions from different domains (SeqMix). We validate SeqDG on the EGTEA and EPIC-KITCHENS-100 datasets. Results on EPIC-KITCHENS-100, show that SeqDG leads to +2.4% relative average improvement in cross-domain action recognition in unseen environments, and on EGTEA the model achieved +0.6% Top-1 accuracy over SOTA in intra-domain action recognition.

Human activities exhibit a strong correlation between actions and the places where these are performed, such as washing something at a sink. More specifically, in daily living environments we may identify particular locations, hereinafter named activity-centric zones , which may afford a set of homogeneous actions. Their knowledge can serve as a prior to favor vision models to recognize human activities. However, the appearance of these zones is scene-specific, limiting the transferability of this prior information to unfamiliar areas and domains. This problem is particularly relevant in egocentric vision, where the environment takes up most of the image, making it even more difficult to separate the action from the context. In this paper, we discuss the importance of decoupling the domain-specific appearance of activity-centric zones from their universal, domain-agnostic representations, and show how the latter can improve the cross-domain transferability of Egocentric Action Recognition (EAR) models. We validate our solution on the EPIC-Kitchens-100 and Argo1M datasets. • We shed light on the side-effects of co-occurrence biases in egocentric vision. • We present EgoZAR, an action recognition model leveraging environmental affordances. • We achieve SOTA performance on EPIC-Kitchens-100 in the domain generalization setting.

Recent transformer-based architectures have shown impressive results in the field of image segmentation. Thanks to their flexibility, they obtain outstanding performance in multiple segmentation tasks, such as semantic and panoptic, under a single unified framework. To achieve such impressive performance, these architectures employ intensive operations and require substantial computational resources, which are often not available, especially on edge devices. To fill this gap, we propose Prototype-based Efficient Mask-Former (PEM), an efficient transformer-based architecture that can operate in multiple segmentation tasks. PEM proposes a novel prototype-based cross-attention which leverages the redundancy of visual features to restrict the computation and improve the efficiency without harming the performance. In addition, PEM introduces an efficient multi-scale feature pyramid network, capable of extracting features that have high semantic content in an efficient way, thanks to the combination of deformable convolutions and context-based self-modulation. We benchmark the proposed PEM architecture on two tasks, semantic and panoptic segmentation, evaluated on two different datasets, Cityscapes and ADE20K. PEM demonstrates outstanding performance on every task and dataset, outperforming task-specific architectures while being comparable and even better than computationally expensive baselines. Code is available at https://github.com/NiccoloCavagnero/PEM.

Human comprehension of a video stream is naturally broad: in a few instants, we are able to understand what is happening, the relevance and relationship of objects, and forecast what will follow in the near future, everything all at once. We believe that - to effectively transfer such an holistic perception to intelligent machines - an important role is played by learning to correlate concepts and to abstract knowledge coming from different tasks, to synergistically exploit them when learning novel skills. To accomplish this, we look for a unified approach to video understanding which combines shared temporal modelling of human actions with minimal overhead, to support multiple downstream tasks and enable cooperation when learning novel skills. We then propose EgoPack, a solution that creates a collection of task perspectives that can be carried across downstream tasks and used as a potential source of additional insights, as a backpack of skills that a robot can carry around and use when needed. We demonstrate the effectiveness and efficiency of our approach on four Ego4D benchmarks, outperforming current state-of-the-art methods. Project webpage: sapeirone.github.io/EgoPack.

Soft robots are gaining popularity thanks to their intrinsic safety to contacts and adaptability. However, the potentially infinite number of Degrees of Freedom makes their modeling a daunting task, and in many cases only an approximated description is available. This challenge makes reinforcement learning (RL) based approaches inefficient when deployed on a realistic scenario, due to the large domain gap between models and the real platform. In this work, we demonstrate, for the first time, how Domain Randomization (DR) can solve this problem by enhancing RL policies for soft robots with: i) robustness w.r.t. unknown dynamics parameters; ii) reduced training times by exploiting drastically simpler dynamic models for learning; iii) better environment exploration, which can lead to exploitation of environmental constraints for optimal performance. Moreover, we introduce a novel algorithmic extension to previous adaptive domain randomization methods for the automatic inference of dynamics parameters for deformable objects. We provide an extensive evaluation in simulation on four different tasks and two soft robot designs, opening interesting perspectives for future research on Reinforcement Learning for closed-loop soft robot control.

In the last decade, most research in Machine Learning contributed to the improvement of existing models, with the aim of increasing the performance of neural networks for the solution of a variety of different tasks. However, such advancements often come at the cost of an increase of model memory and computational requirements. This represents a significant limitation for the deployability of research output in realistic settings, where the cost, the energy consumption, and the complexity of the framework play a crucial role. To solve this issue, the designer should search for models that maximise the performance while limiting its footprint. Typical approaches to reach this goal rely either on manual procedures, which cannot guarantee the optimality of the final design, or upon Neural Architecture Search algorithms to automatise the process, at the expenses of extremely high computational time. This paper provides a solution for the fast identification of a neural network that maximises the model accuracy while preserving size and computational constraints typical of tiny devices. Our approach, named FreeREA, is a custom cell-based evolution NAS algorithm that exploits an optimised combination of training-free metrics to rank architectures during the search, thus without need of model training. Our experiments, carried out on the common benchmarks NAS-Bench-101 and NATS-Bench, demonstrate that i) FreeREA is a fast, efficient, and effective search method for models automatic design; ii) it outperforms State of the Art training-based and training-free techniques in all the datasets and benchmarks considered, and iii) it can easily generalise to constrained scenarios, representing a competitive solution for fast Neural Architecture Search in generic constrained applications. The code is available at https://github.com/NiccoloCavagnero/FreeREA.

Deep Neural Networks (DNNs) enable a wide series of technological advancements, ranging from clinical imaging, to predictive industrial maintenance and autonomous driving. However, recent findings indicate that transient hardware faults may corrupt the models prediction dramatically. For instance, the radiation-induced misprediction probability can be so high to impede a safe deployment of DNNs models at scale, urging the need for efficient and effective hardening solutions. In this work, we propose to tackle the reliability issue both at training and model design time. First, we show that vanilla models are highly affected by transient faults, that can induce a performances drop up to 37%. Hence, we provide three zero-overhead solutions, based on DNN re-design and re-train, that can improve DNNs reliability to transient faults up to one order of magnitude. We complement our work with extensive ablation studies to quantify the gain in performances of each hardening component.

Wearable sensing has emerged as a promising solution for enabling unobtrusive and ergonomic measurements of the human motion. However, the reconstruction performance of these devices strongly depends on the quality and the number of sensors, which are typically limited by wearability and economic constraints. A promising approach to minimize the number of sensors is to exploit dimensionality reduction approaches that fusepriorinformation with insufficient sensing signals, through minimum variance estimation. These methods were successfully used for static hand pose reconstruction, but their translation to motion reconstruction has not been attempted yet. In this work, we propose the usage of functional principal component analysis to decompose multimodal, time-varying motion profiles in terms of linear combinations of basis functions. Functional decomposition enables the estimation of thea prioricovariance matrix, and hence the fusion of scarce and noisy measured data witha prioriinformation. We also consider the problem of identifying which elemental variables to measure as the most informative for a given class of tasks. We applied our method to two different datasets of upper limb motion D1 (joint trajectories) and D2 (joint trajectories + EMG data) considering an optimal set of measures (four joints for D1 out of seven, three joints, and eight EMGs for D2 out of seven and twelve, respectively). We found that our approach enables the reconstruction of upper limb motion with a median error of$0.013 \pm 0.006$rad for D1 (relative median error 0.9%), and$0.038 \pm 0.023$rad and$0.003 \pm 0.002$mV for D2 (relative median error 2.9% and 5.1%, respectively).

The need for users' safety and technology accept-ability has incredibly increased with the deployment of co-bots physically interacting with humans in industrial settings, and for people assistance. A well-studied approach to meet these requirements is to ensure human-like robot motions. Classic solutions for anthropomorphic movement generation usually rely on optimization procedures, which build upon hypotheses devised from neuroscientific literature, or capitalize on learning methods. However, these approaches come with limitations, e.g. limited motion variability or the need for high dimensional datasets. In this work, we present a technique to directly embed human upper limb principal motion modes computed through functional analysis in the robot trajectory optimization. We report on the implementation with manipulators with redundant anthropomorphic kinematic architectures - although dissimilar with respect to the human model used for functional mode extraction - via Cartesian impedance control. In our experiments, we show how human trajectories mapped onto a robotic manipulator still exhibit the main characteristics of human-likeness, e.g. low jerk values. We discuss the results with respect to the state of the art, and their implications for advanced human-robot interaction in industrial co-botics and for human assistance.

Recently, the avenue of adaptable, soft robotic hands has opened simplified opportunities to grasp different items; however, the potential of soft end effectors (SEEs) is still largely unexplored, especially in human-robot interaction. In this paper, we propose, for the first time, a simple touch-based approach to endow a SEE with autonomous grasp sensory-motor primitives, in response to an item passed to the robot by a human (human-to-robot handover). We capitalize on human inspiration and minimalistic sensing, while hand adaptability is exploited to generalize grasp response to different objects. We consider the Pisa/IIT SoftHand (SH), an under-actuated soft anthropomorphic robotic hand, which is mounted on a robotic arm and equipped with Inertial Measurement Units (IMUs) on the fingertips. These sensors detect the accelerations arisen from contact with external items. In response to a contact, the hand pose and closure are planned for grasping, by executing arm motions with hand closure commands. We generate these motions from human wrist poses acquired from a human maneuvering the SH to grasp an object from a table. We obtained 86% of successful grasps, considering many objects passed to the SH in different manners. We also tested our techniques in preliminary experiments, where the robot moved to autonomously grasp objects from a surface. Results are positive and open interesting perspectives for soft robotic manipulation.

An effective robotic wrist represents a key enabling element in robotic manipulation, especially in prosthetics. In this paper, we propose an under-actuated wrist system, which is also adaptable and allows to implement different under-actuation schemes. Our approach leverages upon the idea of soft synergies — in particular the design method of adaptive synergies — as it derives from the field of robot hand design. First we introduce the design principle and its implementation and function in a configurable test bench prototype, which can be used to demonstrate the feasibility of our idea. Furthermore, we report on results from preliminary experiments with humans, aiming to identify the most probable wrist pose during the pre-grasp phase in activities of daily living. Based on these outcomes, we calibrate our wrist prototype accordingly and demonstrate its effectiveness to accomplish grasping and manipulation tasks.