In the context of smart buildings and smart cities, the design of low-cost and privacy-aware solutions for recognizing the presence of humans and their activities is becoming of great interest. Existing solutions exploiting wearables and video-based systems have several drawbacks, such as high cost, low usability, poor portability, and privacy-related issues. Consequently, more ubiquitous and accessible solutions, such as WiFi sensing, became the focus of attention. However, at the current state-of-the-art, WiFi sensing is subject to low accuracy and poor generalization, primarily affected by environmental factors, such as humidity and temperature variations, and furniture position changes. Such is-sues are partially solved at the cost of complex data preprocessing pipelines. In this paper, we present a highly accurate, resource-efficient deep learning-based occupancy detection solution, which is resilient to variations in humidity and temperature. The approach is tested on an extensive benchmark, where people are free to move and the furniture layout does change. In addition, based on a consolidated algorithm of explainable AI, we quantify the importance of the WiFi signal w.r.t. humidity and temperature for the proposed approach. Notably, humidity and temperature can indeed be predicted based on WiFi signals; this promotes the expressivity of the WiFi signal and at the same time the need for a non-linear model to properly deal with it.

Industry 4.0 involves the integration of digital technologies, such as IoT, Big Data, and AI, into manufacturing and industrial processes to increase efficiency and productivity. As these technologies become more interconnected and interdependent, Industry 4.0 systems become more complex, which brings the difficulty of identifying and stopping anomalies that may cause disturbances in the manufacturing process. This paper aims to propose a diffusion-based model for real-time anomaly prediction in Industry 4.0 processes. Using a neuro-symbolic approach, we integrate industrial ontologies in the model, thereby adding formal knowledge on smart manufacturing. Finally, we propose a simple yet effective way of distilling diffusion models through Random Fourier Features for deployment on an embedded system for direct integration into the manufacturing process. To the best of our knowledge, this approach has never been explored before.

We propose an end-to-end solution to address the problem of object localisation in partial scenes, where we aim to estimate the position of an object in an unknown area given only a partial 3D scan of the scene. We propose a novel scene representation to facilitate the geometric reasoning, Directed Spatial Commonsense Graph (D-SCG), a spatial scene graph that is enriched with additional concept nodes from a commonsense knowledge base. Specifically, the nodes of D-SCG represent the scene objects and the edges are their relative positions. Each object node is then connected via different commonsense relationships to a set of concept nodes. With the proposed graph-based scene representation, we estimate the unknown position of the target object using a Graph Neural Network that implements a sparse attentional message passing mechanism. The network first predicts the relative positions between the target object and each visible object by learning a rich representation of the objects via aggregating both the object nodes and the concept nodes in D-SCG. These relative positions then are merged to obtain the final position. We evaluate our method using Partial ScanNet, improving the state-of-the-art by 5.9% in terms of the localisation accuracy at a 8x faster training speed.

We solve object localisation in partial scenes, a new problem of estimating the unknown position of an object (e.g. where is the bag?) given a partial 3D scan of a scene. The proposed solution is based on a novel scene graph model, the Spatial Commonsense Graph (SCG), where objects are the nodes and edges define pairwise distances between them, enriched by concept nodes and relationships from a commonsense knowledge base. This allows SCG to better generalise its spatial inference over unknown 3D scenes. The SCG is used to estimate the unknown position of the target object in two steps: first, we feed the SCG into a novel Proximity Prediction Network, a graph neural network that uses attention to perform distance prediction between the node representing the target object and the nodes representing the observed objects in the SCG; second, we propose a Localisation Module based on circular intersection to estimate the object position using all the predicted pairwise distances in order to be independent of any reference system. We create a new dataset of partially reconstructed scenes to benchmark our method and baselines for object localisation in partial scenes, where our proposed method achieves the best localisation performance.

Retrieving clothes that are worn in social media videos (Instagram, TikTok) is the latest frontier of e-fashion, referred to as "video-to-shop" in the computer vision literature. In this paper, we present MovingFashion, the first publicly available dataset to cope with this challenge. MovingFashion is composed of 14855 social videos, each one of them associated with e-commerce "shop" images where the corresponding clothing items are clearly portrayed. In addition, we present a novel baseline for this scenario, dubbed SEAM Match-RCNN. The model is trained by image-tovideo domain adaptation, allowing the use of video sequences where only their association with a shop image is given, eliminating the need for millions of annotated bounding boxes. SEAM Match-RCNN builds an embedding, where an attention-based weighted sum of few frames (10) of a social video is enough to individuate the correct product within the first 5 retrieved items in a 14K+ shop element gallery with an accuracy of 80%. This provides the best performance on MovingFashion, comparing exhaustively against the related state-of-the-art approaches and alternative baselines1.

This paper presents DOHMO, an embedded computer vision system where multiple sensors, including intelligent cameras, are connected to actuators that regulate illumination and doors. The system aims at assisting elderly and impaired people in co-housing scenarios, in accordance with privacy design principles. The paper provides details of two core elements of the system: The first one is the BOX-IO controller, a fully scalable and customizable hardware and software IoT ecosystem that can collect, control, and monitor data, operational flows and business scenarios, whether indoor or outdoor. The second one is the embedded 3DEverywhere intelligent camera, a device composed of an embedded system that receives input data provided by a 3D/2D camera, analyzes it, and returns the metadata of this analysis. We illustrate how they can be connected and how simple decision mechanisms can be implemented in such a framework. In particular, illumination can be triggered on and off by the detected presence of people, overcoming the limitations of typical sensors, while doors can be opened or closed based on person trajectories in an intelligent manner. To substantiate the proposed system, numerous experiments are performed in a lab and a co-housina scenario.