Orientation (computer vision)

About: Orientation (computer vision) is a research topic. Over the lifetime, 17196 publications have been published within this topic receiving 358181 citations.

Location-based augmented reality on mobile phones

Abstract: The computational capability of mobile phones has been rapidly increasing, to the point where augmented reality has become feasible on cell phones. We present an approach to indoor localization and pose estimation in order to support augmented reality applications on a mobile phone platform. Using the embedded camera, the application localizes the device in a familiar environment and determines its orientation. Once the 6 DOF pose is determined, 3D virtual objects from a database can be projected into the image and displayed for the mobile user. Off-line data acquisition consists of acquiring images at different locations in the environment. The online pose estimation is done by a feature-based matching between the cell phone image and an image selected from the precomputed database using the phone's sensors (accelerometer and magnetometer). The application enables the user both to visualize virtual objects in the camera image and to localize the user in a familiar environment. We describe in detail the process of building the database and the pose estimation algorithm used on the mobile phone. We evaluate the algorithm performance as well as its accuracy in terms of reprojection distance of the 3D virtual objects in the cell phone image.

Registration of three dimensional image data with patient in a projection imaging system

Abstract: A method for determining a translation of a three-dimensional pre-operative image data set to obtain a registration of the three-dimensional image data with a patient positioned in a projection imaging system. In one embodiment the user identifies an initial three-dimensional organ center from projections and extreme contour landmark points of the object on a set of projections. A set of contour points for the image object in each of a plurality of three-dimensional cross-section planes; is obtained and the points projecting nearest to the user-identified landmark points are selected. A three-dimensional grid having a predetermined number of intervals at a predetermined interval spacing centered at the user-identified organ center is defined. The three-dimensional image data contour points as centered onto each grid point are projected for evaluation and selection of the grid point leading to contour points projecting nearest to the user-identified landmark points. This selection leads to the iterative definition of a series of improved estimated three-dimensional organ centers, and associated translation vectors. Registration of a three dimensional image data to the patient positioned in a projection imaging system will allow, among other things, overlay of a visual representation of a pre-operative image object onto a projection image plane that can serve as a visual tool and a surgical navigation aid. In particular, the position and orientation of a medical device can be shown with respect to the three-dimensional image data and thus enable quicker, safer, and less invasive navigation of the medical device to and within an organ of interest.

Development of a real time image based object recognition method for mobile AR-devices

Abstract: In this paper we describe an image based object recognition and tracking method for mobile AR-devices and the correlative process to generate the required data. The object recognition and tracking base on the 3D-geometries of the related objects. Correspondings between live camera images and 3D-models are generated and used to determine the location and orientation of objects in the current scene. The required data for the object recognition is generated from common 3D-CAD-files using a dedicated process model.The object recognition and tracking method as well as the correlative generation process for the needed data are developed within the AR-PDA project. The AR-PDA is a personal digital assistant (e.g. PDA or 3rd generation mobile phone with an integrated camera), which uses AR technology to efficiently support consumers and service forces during their daily tasks.

RADet: Refine Feature Pyramid Network and Multi-Layer Attention Network for Arbitrary-Oriented Object Detection of Remote Sensing Images

Abstract: Object detection has made significant progress in many real-world scenes. Despite this remarkable progress, the common use case of detection in remote sensing images remains challenging even for leading object detectors, due to the complex background, objects with arbitrary orientation, and large difference in scale of objects. In this paper, we propose a novel rotation detector for remote sensing images, mainly inspired by Mask R-CNN, namely RADet. RADet can obtain the rotation bounding box of objects with shape mask predicted by the mask branch, which is a novel, simple and effective way to get the rotation bounding box of objects. Specifically, a refine feature pyramid network is devised with an improved building block constructing top-down feature maps, to solve the problem of large difference in scales. Meanwhile, the position attention network and the channel attention network are jointly explored by modeling the spatial position dependence between global pixels and highlighting the object feature, for detecting small object surrounded by complex background. Extensive experiments on two remote sensing public datasets, DOTA and NWPUVHR -10, show our method to outperform existing leading object detectors in remote sensing field.

System and method for sharing virtual and augmented reality scenes between users and viewers

Abstract: A preferred method for sharing user-generated virtual and augmented reality scenes can include receiving at a server a virtual and/or augmented reality (VAR) scene generated by a user mobile device. Preferably, the VAR scene includes visual data and orientation data, which includes a real orientation of the user mobile device relative to a projection matrix. The preferred method can also include compositing the visual data and the orientation data into a viewable VAR scene; locally storing the viewable VAR scene at the server; and in response to a request received at the server, distributing the processed VAR scene to a viewer mobile device.