Action recognition plays an important role in various applications such as video monitoring, automatic video indexing, crowd analysis, human-machine interaction, smart homes and personal assistive robotics. In this paper, we propose improvements to some methods for human action recognition from videos that work with data represented in the form of skeleton poses. These methods are based on the most widely used techniques for this problem—Graph Convolutional Networks (GCNs), Temporal Convolutional Networks (TCNs) and Recurrent Neural Networks (RNNs). Initially, the paper explores and compares different ways to extract the most relevant spatial and temporal characteristics for a sequence of frames describing an action. Based on this comparative analysis, we show how a TCN type unit can be extended to work even on the characteristics extracted from the spatial domain. To validate our approach, we test it against a benchmark often used for human action recognition problems and we show that our solution obtains comparable results to the state-of-the-art, but with a significant increase in the inference speed.

Robust action recognition methods lie at the cornerstone of Ambient Assisted Living (AAL) systems employing optical devices. Using 3D skeleton joints extracted from depth images taken with time-of-flight (ToF) cameras has been a popular solution for accomplishing these tasks. Though seemingly scarce in terms of information availability compared to its RGB or depth image counterparts, the skeletal representation has proven to be effective in the task of action recognition. This paper explores different interpretations of both the spatial and the temporal dimensions of a sequence of frames describing an action. We show that rather intuitive approaches, often borrowed from other computer vision tasks, can improve accuracy. We report results based on these modifications and propose an architecture that uses temporal convolutions with results comparable to the state of the art.

Space Surveillance and Tracking is a task that requires the development of systems that can accurately discriminate between natural and man-made objects that orbit around Earth. To manage the discrimination between these objects, it is required to analyze a large amount of partially annotated astronomical images collected using a network of on-ground and potentially space-based optical telescopes. Thus, the main objective of this article is to propose a novel architecture that improves the automatic annotation of astronomical images. To achieve this objective, we present a new method for automatic detection and classification of space objects (point-like and streaks) in a supervised manner, given real-world partially annotated images in the FITS (Flexible Image Transport System) format. Results are strongly dependent on the preprocessing techniques applied to the images. Therefore, different techniques were tested including our method for object filtering and bounding box extraction. Based on our relabeling pipeline, we can easily follow how the number of detected objects is gradually increasing after each iteration, achieving a mean average precision of 98%.

The task of human action recognition (HAR) can be found in many computer vision practical applications. Various data modalities have been considered for solving this task, including joint-based skeletal representations which are suitable for real-time applications on platforms with limited computational resources. We propose a spatio-temporal neural network that uses handcrafted geometric features to classify human actions from video data. The proposed deep neural network architecture combines graph convolutional and temporal convolutional layers. The experiments performed on public HAR datasets show that our model obtains results similar to other state-of-the-art methods but has a lower inference time while offering the possibility to obtain an explanation for the classified action.

Background: Whole-slide images (WSIs) are crucial in pathology for digitizing tissue slides, enabling pathologists and AI models to analyze cancer patterns at gigapixel scale. However, their large size incorporates artifacts and non-tissue regions that slow AI processing, consume resources, and introduce errors like false positives. Tissue detection serves as the essential first step in WSI pipelines to focus on relevant areas, but deep learning detection methods require extensive manual annotations. Methods: This study benchmarks four thumbnail-level tissue detection methods—Otsu’s thresholding, K-Means clustering, our novel annotation-free Double-Pass hybrid, and GrandQC’s UNet++ on 3322 TCGA WSIs from nine cancer cohorts, evaluating accuracy, speed, and efficiency. Results: Double-Pass achieved an mIoU of 0.826—very close to the deep learning GrandQC model’s 0.871—while processing slides on a CPU in just 0.203 s per slide, markedly faster than GrandQC’s 2.431 s per slide on the same hardware. As an annotation-free, CPU-optimized method, it therefore enables efficient, scalable thumbnail-level tissue detection on standard workstations. Conclusions: The scalable, annotation-free Double-Pass pipeline reduces computational bottlenecks and facilitates high-throughput WSI preprocessing, enabling faster and more cost-effective integration of AI into clinical pathology and research workflows. Comparing Double-Pass against established methods, this benchmark demonstrates its novelty as a fast, robust and annotation-free alternative to supervised methods.