Hager G, Toyama K. XVision: a portable substrate for real-time vision applications. Comput Vision ImageUnd-kqqfiJ ing 1998;65(1):14--26.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....was maintained; Figure 4: Step Input Joint Position Response Figure 5: Step Input Joint Velocity Response and (c) the camera s image plane remained parallel to the gripper. Each fiducial mark was tracked with a (40 Theta 40) window using Hager s X Vision sum ofsquare difference (SSD) trackers [11]. This allowed us to do tracking in real time at the 30 fps video rate. In all tests, the lens focal length, f , was 12.5 mm: and z was set to 25 cm. The step input was carried out as follows. First, a reference image was defined (Figure 3 Left) by placing a SSD tracker on each fiducial mark. ....

Hager, G.D., Toyama, K., "X vision: A Portable Substrate for Real-Time Vision Applications", Tech Report, Dept of Computer Science, Yale Univ, 1995.

....and changed on the fly. The net effect is a 5 DOF hybrid eye in hand robot with two separate motion controllers and with different servo update rates. Images are acquired with an off the shelf charge coupled device (CCD) camera and digitized by a Sparc 20 installed framegrabber and X Vision [10] was used for image processing. The Sparc is the host computer which handles image data (acquisition and processing) and issues ASCII encoded gantry and PTU motion commands via serial communications. OH AND ALLEN: VISUAL SERVOING BY PARTITIONING DEGREES OF FREEDOM 3 Fig. 3. Robot and hardware ....

G. D. Hager and K. Toyama, "X vision: A portable substrate for real-time vision applications," Dept. Comput. Sci., Yale Univ., Tech. Rep., 1995.

....visual feature tracking optimized to be simple to configure at the user level. Its performance relies on using only small sub windows for tracking and representing features in canonical configuration which allow faster computation of the actual feature position. More information can be found in [11, 6, 21]. It is free available for non commercial use. Human Interface Technology Lab. www.hitl.washington.edu) ARToolkit: provides a portable computer vision tracking library. It is a software library that can be used to calculate camera position and orientation relative to physical markers in real ....

G.D. Hager and K. Toyama. XVision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 69(1):23--37, January 1998.

....system does not bene t from a stereo setup and currently uses only one of the two cameras. We still stick to a stereo setup which will be required for future use of natural landmarks, as 3D reconstruction is not possible with a monocular setup. 2.2. 2D feature tracking We use the XVision library [6] for 2D tracking. XVision allows for fast tracking and e cient computation of feature locations by using canonical feature representations. In our system, we use corners as tracking primitives. In XVision, each corner is composed of two lines. While the XVision line tracking primitive is very ....

Hager GD, Toyama K. XVision: a portable substrate for real-time vision applications. Computer Vision and Image Understanding 1998;69(1):23}37.

No context found.

G. D. Hager and K. Toyama. Xvision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 69(1):23-- 37, 1996.

No context found.

# G.D. Hager and K. Toyama, "XVision: A Portable Substrate for Real-Time Vision Applications," Computer Vision and Image Understanding, vol. 69, no. 1, pp. 23-37, 1998.

No context found.

G. D. Hager and K. Toyama. Xvision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 69(1):23-- 37, 1996.

No context found.

G. D. Hager and K. Toyama. Xvision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 69(1):23--37, 1996.

No context found.

G. D. Hager and K. Toyama, "XVision: A portable substrate for real-time vision applications, " Computer Vision and Image Understanding, vol. 69, no. 1, 1998.

....illumination compensation, and outlier detection. All experiments were performed on live video sequences by an SGI Indy equipped with a 175 Mhz R4400 SC processor and VINO image acquisition system. 5. 1 Implementation We have implemented the methods described above within the X Vision environment [41]. The implemented system incorporates all of the linear motion models described in Section 2, nonorthonormal illumination bases as described in Section 3, and outlier rejection using the algorithm described in Section 4. The image warping required to support the algorithm is implemented by ....

....shearing the region using bilinear interpolation. The resolution of the region is then reduced by averaging neighboring pixels. Spatial and temporal derivatives are computed by applying Prewitt operators on the reduced scale images. More details on this level of the implementation can be found in [41]. In these experiments, the algorithm is initialized by interactively selecting a region to track in a live video stream. The algorithm immediately acquires the selected region as the reference template and performs tracking on all subsequent images of the stream. When an illumination basis is ....

[Article contains additional citation context not shown here]

 G.D. Hager and K. Toyama, "XVision: A Portable Substrate for Real-Time Vision Applications," Computer Vision and Image Understanding, vol. 69, no. 1, pp. 23-37, 1998.

....points on the object by slight modification of the objective function. The optimization could also be easily extended to perform a robust optimization step using IRLS methods [42] making it yet more robust to outliers. We are currently implementing a version of the algorithm within the XVision [43] environment for use in robotic applications, as well as augmented and virtual reality. An initial implementation described in [44] has shown that, by combining efficient local tracking with efficient pose estimation, it is relatively simple to construct real time object tracking system which run ....

G. D. Hager and K. Toyama, "XVision: A portable substrate for real-time vision applications, " Computer Vision and Image Understanding, 1998. To Appean, Jan. 1998.

....processor and VINO image acquisition system. X Y Rotation Scale Aspect Ratio Shear Figure 2: The columns of the motion Jacobian matrix for the planar target and their geometric interpretations. 5. 1 Implementation We have implemented the methods described above within the X Vision environment [23]. The implemented system incorporates all of the linear motion models described in Section 2, non orthonormal illumination bases as described in Section 3, and outlier rejection using the algorithm described in Section 4. The image warping required to support the algorithm is implemented by ....

....scaled and sheared using a bilinear interpolation. The resolution of the region is then reduced by averaging neighboring pixels. Spatial and temporal derivatives are computed by applying Prewitt operators on the reduced scale images. More details on this level of the implementation can be found in [23]. Timings of the algorithm 2 indicate that it can perform frame rate (30 Hz) tracking of image regions of up to 100 Theta 100 pixels at one half resolution undergoing affine distortions and illumination changes. Similar performance has been achieved on a 120Mhz Pentium processor and 70 Mhz Sun ....

[Article contains additional citation context not shown here]

G. D. Hager and K. Toyama. XVision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 1996. In Press.

No context found.

Hager, G., and Toyama, K. 1996a. XVision: A portable substrate for real-time vision applications. In Proc.

....image and computes a residue value. It returns an output set whose elements are configurations that correspond to affine transformations of the template which match with the current camera image. It fails if the computed sum of squared difference (SSD) residue is above a given threshold (see [17]) The residue of an SSD computation is the actual pixel by pixel sum of squared differences. Variations of this tracker may perform the same function with a subset of the affine transformations. ffl An SSD based multi pattern tracker performs as above, but it tracks successfully if the current ....

G. Hager and K. Toyama. XVision: A portable substrate for real-time vision applications. Computer Vision and Image Understanding, 1998.

....temporal filtering, or ad hoc methods for handling specific types of visual perturbations. For instance, distractions objects which are close to the target both in appearance and state can be handled in several ways. One method is to consider only a small set of states surrounding the target [15, 47]. This requires some predictability in the target trajectory but eliminates the need to examine the entire image. Another way to handle distraction is through foveation, effectively blurring the image region around the target [6, 37] Finally, a tracking system may filter output estimates based on ....

G. Hager and K. Toyama. XVision: A portable substrate for real-time vision applications. In Proc. European Conf. on Computer Vision, volume 2, pages 507--512, Cambridge, UK, 1996.

No context found.

G. Hager and K. Toyama. Xvision: A portable substrate for real-time vision applications. In Proc. ECCV, 1996.

No context found.

Hager G, Toyama K. XVision: a portable substrate for real-time vision applications. Comput Vision ImageUnd-kqqfiJ ing 1998;65(1):14--26.

No context found.

G. Hager and K. Toyama, "X vision: A portable substrate for real-time vision applications," Comput. Vision and Image Understanding, vol. 69, no. 1, pp. 23--27, Jan. 1996.

No context found.

Hager, G. and Toyama, K., "X Vision: A Portable Substrate for Real-Time Vision Applications," Computer Vision and Image Understanding, Vol. 69, No. 1, pp 23-27, January 1996.

No context found.

G. Hager and T. K., "X vision: A portable substrate for real-time vision applications." Comuter Vision and Image understanding, vol. 69, no. 1, pp. 23--37, 1998.

No context found.

G.D. Hager and K. Toyama, XVision: A portable substrate for real-time vision applications, Comput Vision Image Understand 69:(1) (1998), 23--37.

No context found.

G. D. Hager and K. Toyama, "X Vision: a portable substrate for real-time vision applications," Computer Vision and Image Understanding, vol. 69, pp. 23--37, Jan. 1998.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC