Image reconstruction for positron emission tomography (PET) is a challenging inverse problem due to its inherent ill-posedness and Poisson counting noise, particularly in lowdose acquisitions. In this work, we introduce a novel Bayesian PET reconstruction framework that leverages a flow-matching generative prior learned from high-quality PET images. By incorporating this prior, we formulate PET reconstruction as a Bayesian inference problem and solve it via latent-space posterior sampling using the unadjusted Langevin algorithm. We evaluate the proposed method through numerical experiments and benchmark it against classical reconstruction methods and state-of-theart deep learning-based baselines. The proposed method achieves improved reconstruction performance while enabling posterior uncertainty quantification.
In this paper, we present a hand pose based gesture representation that can effectively classify both static and dynamic gestures. In contrast to resource and data intensive deep learning models, we postulate that hand pose alone can be used to extract compact yet discriminative features that are suitable for most applications that require real-time gesture recognition with minimal computational overhead. Building on the robustness of modern hand pose estimation frameworks and the expressive power of neural networks, we extract a fine-grained gesture description by decomposing gestures in a set of defined attributes. Our approach is highly capable of generalizing to unseen data and unseen classes, as shown by our experiments in the context of the BonsAPPs Challenge use-case.
Surround view systems are becoming more and more important in Autonomous Driving and Advanced Driver Assistance Systems where multiple perception tasks have been developed in order to solve the needs in this field. However, there is a lack of literature for multi-object trackers that deals with multiple cameras in automotive setups. In this paper, we propose a system to tackle multi-object tracking with inter-camera object identification using fisheye cameras covering a 360º view in real time. We provide results from a custom ground truth annotated sequence using CARLA simulator.
Helicoverpa Armigera, or cotton bollworm, is a serious insect pest of cotton crops that threatens the yield and the quality of lint. The timely knowledge of the presence of the insects in the field is crucial for effective farm interventions. Meteo-climatic and vegetation conditions have been identified as key drivers of crop pest abundance. In this work, we applied an interpretable classifier, i.e., Explainable Boosting Machine, which uses earth observation vegetation indices, numerical weather predictions and insect trap catches to predict the onset of bollworm harmfulness in cotton fields in Greece. The glass-box nature of our approach provides significant insight on the main drivers of the model and the interactions among them. Model interpretability adds to the trustworthiness of our approach and therefore its potential for rapid uptake and context-based implementation in operational farm management scenarios. Our results are satisfactory and the importance of drivers, through our analysis on global and local explainability, is in accordance with the literature.
Automatic detection of natural disasters and incidents has become more important as a tool for fast response. There have been many studies to detect incidents using still images and text. However, the number of approaches that exploit temporal information is rather limited. One of the main reasons for this is that a diverse video dataset with various incident types does not exist. To address this need, in this paper we present a video dataset — Video Dataset of Incidents, VIDI — that contains 4,534 video clips corresponding to 43 incident categories. Each incident class has around 100 videos with a duration of ten seconds on average. To increase diversity, the videos have been searched in several languages. To assess the performance of the recent state-of-the-art approaches, Vision Transformer and TimeSformer, as well as to explore the contribution of video-based information for incident classification, we performed benchmark experiments on the VIDI and Incidents Dataset. We have shown that the recent methods improve the incident classification accuracy. We have found that employing video data is very beneficial for the task. By using the video data, the top-1 accuracy is increased to 76.56% from 67.37%, which was obtained using a single frame. VIDI will be made publicly available. Additional materials can be found at the following link: https://github.com/vididataset/VIDI
