Citation Context

... Axes are also displayed for each object as well as a pink label. Top image misses two objects; all are found in the bottom one. While there are real-time feature trackers that can run on a cellphone =-=[15]-=-, they usually operate on very small images (e.g., 120x160) and with very few features. Systems comparable to ours [30] typically take over 1 second per image. We were able to run ORB with 640 × 480 r...

Citation Context

...LAM systems based on a set of three dimensionalinformationfilterswhichcanestimatetheposition of arbitrarily distant features. Very recently, a similar method was briefly described by Klein and Murray =-=[14]-=-. Compared to our approach, their filters are applied on every frame, are one-dimensional and are only used for data association. Ultimately, we would like to update a set of partially initialised 3D ...

Citation Context

...ffreedom (6DOF) at real-time update rates. Previous research primarily uses computer vision techniques to compute the localization. Such 6DOF localization works on mobile devices for small workspaces =-=[23, 12]-=-, but scaling such techniques to larger environments, such as an entire city, is challenging. 1.1 Related Work Some recent work formulates the city-scale localization as an image retrieval problem: as...

Citation Context

....g. in the movie industry, and it has been successfully used to build 3D models from both photo collections, and from video [24]. Another technology that uses SfM as its back-end is augmented reality =-=[16]-=-. An overwhelming majority of image sensors sold today are of CMOS type: nearly all mobile video recording devices, and most compact cameras have them. In contrast to the classical CCD sensors, which ...

Citation Context

...em set, the mobile space introduces additional important constraints on computation capabilities and power usage. While there have been some investigation into the mobile vision space in recent years =-=[20]-=-, it is just now becoming a focus. Previous vision benchmark efforts have either looked at computer vision components as a part of a larger benchmark or provided benchmarks with basic timing analysis....

Citation Context

... a wide range of applications, such as augmenting ansentire book with digital interactions or linking city-wide moviesposters or leaflets with multimedia trailers and advisements.sSeveral MAR systems =-=[10, 11, 12, 13, 14, 15, 19]-=- have beensproposed recently, such as Wagner et al.’s pose tracking [10, 11,s12], Klein and Murray’s parallel tracking and mapping [13] andsTa et al.’s SURFTrac [14]. Despite of their success in real-...

Citation Context

... posters can be used as distinctive regions. In addition to estimating absolute locations, it is also possible to compute relative locations using visual odometry, e.g. by tracking features over time =-=[7]-=- or using inertial sensors to estimate the relative movement. Vision-based indoor navigation is one promising candidate technology since it has the potential to work without an additional (costly) inf...

Citation Context

...do not directly handle the RS but circumvent its effects by first undistorting the frames with methods as proposed in [15] and subsequently applying a state-of-the-art GS SfM method. Klein and Murray =-=[12]-=-, for example, port their parallel tracking and mapping (PTAM) framework [11] to an Apple iPhone 3GS. They first estimate a constant camera velocity between two consecutive frames which is then used t...

Citation Context

...knowledge, all existing approaches rely exclusively on visual data, and do not use any inertial sensors. For example, [14], [15] present algorithms for bundle adjustment with rolling-shutter cameras, =-=[16]-=- describes an implementation of the wellknown PTAM algorithm [17] using a rolling-shutter sensor, while [18]–[20] propose methods for estimating the motion of objects by exploiting the rolling shutter...