
Computer vision focuses on the automated extraction, analysis, and interpretation of visual information from images and videos. This discipline integrates image processing, pattern recognition, and machine learning to address tasks such as object detection, segmentation, recognition, and 3D reconstruction. Research explores advancements in convolutional neural networks, optical flow, depth estimation, and scene understanding, enabling applications across robotics, surveillance, diagnostica per immagini, and autonomous systems. The following collection presents recent scholarly publications and patented technologies that advance metodologie and applications within computer vision:
Questa è la nostra ultima selezione di pubblicazioni e brevetti in lingua inglese sulla Computer Vision, tra numerose riviste scientifiche online, classificate e focalizzate su segmentazione di immagini, rilevamento di oggetti, estrazione di caratteristiche, rete neurale convoluzionale, classificazione di immagini, flusso ottico, visione stereo, stima della profondità, riconoscimento di immagini, comprensione della scena, ricostruzione dell'immagine, miglioramento dell'immagine, riconoscimento di pattern, riconoscimento facciale, traccia del movimento, ricostruzione 3D, registrazione dell'immagine, Computer Vision, segmentazione semantica, segmentazione di istanze, SLAM visivo, stitching di immagini, analisi di texture, rilevamento di bordi, tracciamento di oggetti, super-risoluzione, denoising di immagini, analisi di video, rilevamento di anomalie e didascalie di immagini.
A method to measure properties of a forest
Patent published on the 2026-10-18 in FI under Ref FI20255344 by OY ARBONAUT LTD [FI] (LeppÄnen Vesa [fi], Jeskanen Jeremias [fi], Sener Ecre [fi])
Abstract: A method of determining properties of forest is disclosed. Accurate properties of forest are important for management of forest, forest operations and forest protection. Producing cross-section images of forest allows image recognition techniques to be used in identification of trees (40, 50). Machine learning methods may be used as one form of image recognition techniques. Tree attributes, like size, height or species may be identified for the identified trees. Having overlapping area in the cr[...]
Our summary: A method for measuring forest properties is disclosed. Cross-section images facilitate tree identification using image recognition techniques. Machine learning aids in recognizing tree attributes and managing duplicate identifications.
forest measurement, image recognition, machine learning, tree attributes
Patent
Water injection performance anomaly detection and trends forecasting using machine learning algorithms
Patent published on the 2026-10-15 in US under Ref US20260318198 by SAUDI ARABIAN OIL CO [SA] (Hommadi Tariq Ali [sa], Momtan Bayan A [sa])
Abstract: A method that includes obtaining fluid injection data from a fluid injection operation performed on a reservoir, the fluid injection operation controlled by fluid injection parameters including a fluid injection rate. The method further includes determining a fluid injection performance record based on the fluid injection data, identifying a plurality of breakpoints of the fluid injection performance record using a first machine learning algorithm and determining, using a second machine learning[...]
Our summary: The method detects anomalies in water injection performance using machine learning. It identifies breakpoints in fluid injection records and forecasts future performance. Updated parameters are determined to optimize future fluid injection operations.
anomaly detection, machine learning, fluid injection, performance forecasting
Patent
System and method for offline 3d object detection in traffic scenes using lidar data with multiple resolutions
Patent published on the 2026-10-15 in US under Ref US20260316782 by TORC ROBOTICS INC [US] (Li Dalong [us], Shoaib Syed Taha [us], Zhu Xinxiang [us], Badgett Marcus [us], Fei Juncong [de])
Abstract: A system of an autonomous vehicle is described. The autonomous vehicle is configured to collect point cloud data associated with a current time, perform a dynamic object detection on the point cloud data to identify at least one dynamic object, and apply a dynamic object weight to the at least one dynamic object. The autonomous vehicle is further configured to retrieve point cloud data associated with at least one additional period of time, aggregate the point cloud data associated with the curr[...]
Our summary: The system detects dynamic and static objects in traffic scenes using lidar data. It aggregates point cloud data from multiple time periods for comprehensive analysis. The vehicle applies weights to identified objects to enhance detection accuracy.
Lidar, 3D object detection, autonomous vehicle, point cloud data
Patent
System and method for detection of objects within receptacles
Patent published on the 2026-10-15 in WO under Ref WO2026217872 by KOONTZ BRENNAN [US] (Koontz Brennan [us])
Abstract: An object detection system includes a waste collection system configured to collect and store waste from one or more receptacles. A sensor system is configured to capture data indicative of one or more defined objects within the one or more receptacles. A computing system includes one or more processors and one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, configure the computing system to perform operations[...]
Our summary: The system detects objects within receptacles using a sensor system. It captures data to identify defined objects present in the waste. The computing system processes this data and generates control outputs based on the findings.
object detection, waste collection, sensor system, computing system
Patent
System and method for analyzing content using artificial intelligence
Patent published on the 2026-10-15 in US under Ref US20260320449 by XOVATION SOLUTIONS LLC [US] (Pourdastan Mahasti [us], Lease David Richard [us])
Abstract: [0000] A system and method for analyzing multimedia content using an agent-orchestrated artificial intelligence architecture is disclosed. The system receives video, audio, image, and text inputs from a user device and performs preliminary feature extraction. A server-based metadata extraction engine processes the input by qualifying modality types, extracting features using machine-learning and language-model techniques, and initiating an agent-based research process. Multiple research agents a[...]
Our summary: The system analyzes multimedia content through an AI architecture. It extracts features from various input modalities using machine learning techniques. An orchestration module coordinates research agents to refine results into structured metadata.
Artificial Intelligence, Multimedia Analysis, Machine Learning, Metadata Extraction
Patent
System and method for beaconless adaptive optics by means of a multi-plane sensor and convolutional neural network
Patent published on the 2026-10-15 in US under Ref US20260323167 by THE GOVERNMENT OF THE US SECRETARY OF THE NAVY [US] (Dicomo Gregory P [us], Witt George A [us])
Abstract: [0000] An apparatus includes a laser emitting laser light. The apparatus includes a wavefront corrector. The wavefront corrector directs the laser light through a distorting optical medium to a scattering object. The scattering object scatters the distorted light from the distorted spot back through the distorting optical medium to impinge upon the wavefront corrector. The apparatus includes a plurality of imaging cameras including a focal plane camera and at least one of a post-focus camera and[...]
Our summary: The system uses laser light and a wavefront corrector to manage distortions from an optical medium. Imaging cameras capture scattered light and generate image data. A convolutional neural network processes this data to adjust the wavefront corrector for improved image quality.
adaptive optics, multi-plane sensor, convolutional neural network, wavefront corrector
Patent
evidence from China and the European Union carbon markets
Published on 2026-10-14 by @OXFORD
Abstract: AbstractAgainst the backdrop of accelerating global carbon neutrality efforts, carbon markets have become important policy instruments for achieving emission reduction targets. However, their complex pricing mechanisms and pronounced nonlinear and nonstationary characteristics continue to challenge accurate forecasting. To address these issues, this study develops a hybrid forecasting framework integrating Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), a convolutio[...]
Our summary: This study develops a hybrid forecasting framework for carbon markets using CEEMDAN, CNN, and BiGRU. It improves forecasting accuracy and stability while analyzing regional differences in influencing factors. A dynamic risk warning module identifies abnormal price movements, aiding market participants and regulators.
carbon markets, forecasting, CEEMDAN, CNN-BiGRU
Publication
Adaptive parallel multi-representation MiniROCKET for time series classification
Published on 2026-10-05 by @OXFORD
Abstract: AbstractTime series classification has achieved strong performance through transformation-based methods such as MiniROCKET, which rely on fast random convolutional feature extraction from raw temporal signals. However, representing time series data in a single domain limits the ability to capture complex temporal–spectral and structural characteristics in diverse real-world datasets. In this study, we propose ManyROCKET, an adaptive parallel MiniROCKET framework that integrates multiple comple[...]
Our summary: ManyROCKET is an adaptive parallel framework for time series classification that integrates multiple signal representations. It utilizes seven established transformations to enhance feature extraction and classification performance. Extensive experiments show that ManyROCKET outperforms the standard MiniROCKET baseline across diverse datasets.
time series classification, MiniROCKET, multi-representation, adaptive framework
Publication

















