Serwis Infona wykorzystuje pliki cookies (ciasteczka). Są to wartości tekstowe, zapamiętywane przez przeglądarkę na urządzeniu użytkownika. Nasz serwis ma dostęp do tych wartości oraz wykorzystuje je do zapamiętania danych dotyczących użytkownika, takich jak np. ustawienia (typu widok ekranu, wybór języka interfejsu), zapamiętanie zalogowania. Korzystanie z serwisu Infona oznacza zgodę na zapis informacji i ich wykorzystanie dla celów korzytania z serwisu. Więcej informacji można znaleźć w Polityce prywatności oraz Regulaminie serwisu. Zamknięcie tego okienka potwierdza zapoznanie się z informacją o plikach cookies, akceptację polityki prywatności i regulaminu oraz sposobu wykorzystywania plików cookies w serwisie. Możesz zmienić ustawienia obsługi cookies w swojej przeglądarce.
Human action recognition has been studied in many fields including computer vision and sensor networks using inertial sensors. However, there are limitations such as spatial constraints, occlusions in images, sensor unreliability, and the inconvenience of users. In order to solve these problems we suggest a sensor fusion method for human action recognition exploiting RGB images from a single fixed...
In this work we propose a new CNN+LSTM architecture for camera pose regression for indoor and outdoor scenes. CNNs allow us to learn suitable feature representations for localization that are robust against motion blur and illumination changes. We make use of LSTM units on the CNN output, which play the role of a structured dimensionality reduction on the feature vector, leading to drastic improvements...
Low-cost consumer depth cameras and deep learning have enabled reasonable 3D hand pose estimation from single depth images. In this paper, we present an approach that estimates 3D hand pose from regular RGB images. This task has far more ambiguities due to the missing depth information. To this end, we propose a deep network that learns a network-implicit 3D articulation prior. Together with detected...
Monocular Visual Odometry (MVO) estimates the camera position and orientation, based on images generated by a single camera. In this paper a new sparse MVO system for camera equipped vehicles is proposed. Three view cyclic Perspective-n-Point with adaptive threshold is used for camera pose estimation, perspective image transformations are used to improve tracking, and a multi-attribute cost function...
Diminished reality (DR) enables us to see through real objects occluding some areas in our field of view. This interactive display has various applications, such as see-through vision to visualize invisible areas, work area visualization in surgery and landscape simulation. In this paper, we propose two underlying problems in see-through vision, in which hidden areas are observed in real time. First,...
Diminished reality (DR) is a technique to remove undesirable objects from a video stream in real time. DR methods calculate a user's camera pose using vision- or sensor-based approaches to recover and overlay a background image to the camera view. Relying on 6DoF camera registration methods, DR results are often ruined due to misregistration. To solve this problem, we propose a registration framework...
In the context of pedestrian navigation, urban environment constitutes a challenging area for both localization and Augmented Reality (AR). In order to display 3D Geographic Information System (GIS) content in AR and to qualify them, we propose to fuse the pose estimated using vision thanks to a precisely known 3D urban furniture model with rotation estimated from inertial and magnetic measurements...
To robustly estimate the pose, classical methods assume some geometrical and temporal assumptions (SfM: Structure from Motion, SLAM: Simultaneous Localization and mapping). These approaches take a pair of images as input and establish correspondences based on global strategy (using the whole image information) or sparse strategy (using key-points features). These correspondences allow solving a set...
Keypoint extraction and matching has been widely studied by the computer vision community, mostly focused on pinhole camera models. In this paper we perform a comparative analysis of four keypoint extraction algorithms applied to full spherical images, particularly in the context of pose estimation. Two of the methods chosen for the comparative study, namely A-KAZE and ASIFT, have been designed considering...
First-person videos (FPVs) captured by wearable cameras have undesired shakiness because of fast changing views. When existing video stabilization techniques are applied, FPVs are transformed into cinematographic videos, losing the First-person motion information (FPMI) such as the recorder's interests and actions. We propose a system that can enhance viewability of FPVs by stabilizing them while...
Camera pose estimation is a fundamental problem of Augmented Reality and 3D reconstruction systems. Recently, despite the new better performing direct methods being developed, state-of-the-art methods are still estimating erroneous poses due to sensor noise, environmental conditions and challenging trajectories. Adding a back-end mapping process, SLAM systems achieve better performance and are more...
This paper presents a novel vision servoing approach using depth maps to perform robotic motion task with field of view (FOV) constraint. The vision servoing scheme relies on the depth information available from an Red Green Blue-Depth (RGB-D) camera. With respect to the previous approaches, the proposed vision servoing approach has the advantage as follow: First, it does not require the estimation...
3D object selection and manipulation is one of the essential features for any augmented reality (AR) system. However, distant object selection and manipulation still suffer from lack of accuracy and precision. This paper introduces an alternate 3D interaction technique for selection and manipulation distant 3D object in in immersive video see-through AR. The proposed interaction technique offers a...
Monocular visual odometry algorithm has been widely used to estimate the pose of aerial robots in GPS denied environments. However, the pure visual system usually has poor robustness in large scale environments. This paper presents a pose estimation algorithm which fuses monocular visual and inertial data using the monocular ORB-SLAM algorithm as the visual framework. Firstly, the scale estimation...
In this paper, we develop a modified differential Structure from Motion (SfM) algorithm that can estimate relative pose from two consecutive frames despite of Rolling Shutter (RS) artifacts. In particular, we show that under constant velocity assumption, the errors induced by the rolling shutter effect can be easily rectified by a linear scaling operation on each optical flow. We further propose a...
Feature extraction and matching are two crucial components in person Re-Identification (ReID). The large pose deformations and the complex view variations exhibited by the captured person images significantly increase the difficulty of learning and matching of the features from person images. To overcome these difficulties, in this work we propose a Pose-driven Deep Convolutional (PDC) model to learn...
We present our design of a real-time vision-based landing pad detection and pose estimation for many Unmanned Aerial Vehicles(UAV) and implementation on Raspberry Pi. We describe the vision algorithm for precise landing pad detection and recognition and estimate the position and orientation of the UAV relative to the landing pad. The vision algorithm is robust, accurate, and computationally inexpensive...
This paper addresses the problem of estimating the alignment pose between two models by using structure-specific local descriptors which are generated by combining 2D image data and 3D contextual shape data. The 2D texture information is represented by a robust SIFT descriptor, and the geometric information is represented by a histogram supported by the orders of curvature and angles between normal...
Head pose estimation from camera images is a computational problem that may influence many sociological, cognitive, interaction and marketing researches. It is especially crucial in the process of visual gaze estimation which accuracy depends not only on eye region analysis, but head inferring as well. Presented method exploits a 3d head model for a user head pose estimation as it outperforms, in...
The paper presents an approach to localize human body joints in 3D coordinates based on a single low resolution depth image. First a framework to generate a database of 80k realistic depth images from a 3D body model is described. Then data preprocessing and normalization procedure, and DNN and MLP artificial neural networks architectures and training are presented. The robustness against camera distance...
Podaj zakres dat dla filtrowania wyświetlonych wyników. Możesz podać datę początkową, końcową lub obie daty. Daty możesz wpisać ręcznie lub wybrać za pomocą kalendarza.