RETIMAC: REal-TIme Motion Analysis Chip

Nesi, Paolo; Innocenti, Fabrizio; Pezzati, P.

doi:10.1109/82.664242

Cited by 5 publications

(10 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, algorithms for optical flow estimation can be profitable when applied [1,4,9,11], such as the interference techniques [28]. These methods are based on optical systems producing interference patterns from a negative image of two equally spaced particle images.…”

Section: (Iii) High Densitymentioning

confidence: 98%

“…In the literature, several main approaches for motion estimation in a regular grid of the image can be identified: spatio-temporal filtering-based, block-matching, pel-recursive, gradient-based, corner tracking, and line tracking approaches [3,11]. Only some of these approaches can be profitably used for motion estimation of fluid flow, thus for the building of vision-based velocimetries instead of using pressure-based tools that are intrusive.…”

Section: Introductionmentioning

confidence: 99%

“…apparent velocity). This is usually called ''velocity field'' or ''motion field'' [1][2][3]. Most of the techniques adopted for evaluating velocity fields consider changes in the image brightness features/areas.…”

Section: Introductionmentioning

confidence: 99%

“…Motion estimation is very useful in important areas: motion-compensated image sequence processing, dynamic scene analysis, and for measuring and tracking system [3,[5][6][7].…”

Section: Introductionmentioning

confidence: 99%

“…In the literature, there exists some motion estimation architectures for motion analysis in real-time e.g. [8][9][10][11].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Vision-Based Particle Tracking Velocimetry

et al. 2001

Self Cite

View full text Add to dashboard Cite

Section: (Iii) High Densitymentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…Motion estimation is very useful in important areas: motion-compensated image sequence processing, dynamic scene analysis, and for measuring and tracking system [3,[5][6][7].…”

Section: Introductionmentioning

confidence: 99%

“…In the literature, there exists some motion estimation architectures for motion analysis in real-time e.g. [8][9][10][11].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Vision-Based Particle Tracking Velocimetry

et al. 2001

Self Cite

View full text Add to dashboard Cite

Analog Integrated 2-D Optical Flow Sensor

Stocker

2005

Analog Integr Circ Sig Process

View full text Add to dashboard Cite

I present a new focal-plane analog very-large-scale-integrated (aVLSI) sensor that estimates optical flow in two visual dimensions. Its computational architecture consists of a two-layer network of locally connected motion units that collectively estimate the optimal optical flow field. The applied gradient-based optical flow model assumes visual motion to be translational and smooth, and is formulated as a convex optimization problem. The model also guarantees that the estimation problem is well-posed regardless of the visual input by imposing a bias towards a preferred motion under ambiguous or noisy visual conditions. Model parameters can be globally adjusted, leading to a rich output behavior. Varying the smoothness strength, for example, can provide a continuous spectrum of motion estimates, ranging from normal to global optical flow. The non-linear network conductances improve the resulting optical flow estimate because they reduce spatial smoothing across large velocity differences and minimize the bias for reliable stimuli. Extended characterization and recorded optical flow fields from a 30 × 30 array prototype sensor demonstrate the validity of the optical flow model and the robustness and functionality of the computational architecture and its implementation.The ability to estimate motion using visual information is important for any natural and artificial agent behaving in a dynamical visual environment. Knowing the relative motions between different objects as well as between objects and the agent is crucial for a cognitive perception of the environment and thus a prerequisite for intelligent behavior. However, the demand for real-time processing and the limited resources available on freely behaving agents impose severe constraints that require efficient computational systems in terms of processing speed, energy consumption, and physical dimensions. These requirements favor parallel computational architectures. Implementations of such architectures become particularly appealing when image sensing and motion estimation circuitry can be combined within a single sensor. Furthermore, a sensor architecture that consists of a topographically uniform array of identical processing units (pixels) has the advantage that processing power scales with array size, thus keeping processing speed independent of spatial array resolution. Analog VLSI circuits are particularly suited for such architectures because they require significantly less power and silicon area than digital circuits for computational tasks of comparable complexity [1]. Furthermore, timecontinuous analog processing matches the continuous nature of visual motion information. Temporal aliasing artifacts do not occur while they can be a significant problem in clocked, sequential circuit implementations, in particular when limited to low frame-rates [2].Visual information is -in general -locally ambiguous and noisy, and often does not allow a unique estimation of visual motion. A prominent ambiguity, commonly referred to as the "aperture problem" [3]...

show abstract

Dynamics of errors in 3D motion estimation and implications for strain-tensor imaging in acoustic elastography

Bilgen

2000

Phys. Med. Biol.

View full text Add to dashboard Cite

For the purpose of quantifying the noise in acoustic elastography, a displacement covariance matrix is derived analytically for the cross-correlation based 3D motion estimator. Static deformation induced in tissue from an external mechanical source is represented by a second-order strain tensor. A generalized 3D model is introduced for the ultrasonic echo signals. The components of the covariance matrix are related to the variances of the displacement errors and the errors made in estimating the elements of the strain tensor. The results are combined to investigate the dependences of these errors on the experimental and signal-processing parameters as well as to determine the effects of one strain component on the estimation of the other. The expressions are evaluated for special cases of axial strain estimation in the presence of axial, axial-shear and lateral-shear type deformations in 2D. The signals are shown to decorrelate with any of these deformations, with strengths depending on the reorganization and interaction of tissue scatterers with the ultrasonic point spread function following the deformation. Conditions that favour the improvements in motion estimation performance are discussed, and advantages gained by signal companding and pulse compression are illustrated.

show abstract

RETIMAC: REal-TIme Motion Analysis Chip

Cited by 5 publications

References 54 publications

A Vision-Based Particle Tracking Velocimetry

A Vision-Based Particle Tracking Velocimetry

Analog Integrated 2-D Optical Flow Sensor

Dynamics of errors in 3D motion estimation and implications for strain-tensor imaging in acoustic elastography

Contact Info

Product

Resources

About