Mean Shift Paper by Comaniciu and Meer Presentation by Carlo - PowerPoint PPT Presentation

Sep 03, 2023 •19 likes •149 views

Mean Shift Paper by Comaniciu and Meer Presentation by Carlo Lopez-Tello What is the Mean Shift Algorithm? A method of finding peaks (modes) in a probability distribution Works without assuming any underlying structure in the

Mean Shift Paper by Comaniciu and Meer Presentation by Carlo Lopez-Tello
What is the Mean Shift Algorithm? ● A method of finding peaks (modes) in a probability distribution ● Works without assuming any underlying structure in the distribution ● Works on multimodal distributions ● Works without assuming the number of modes
Why do we care about modes? ● Given a data set we can assume that it was sampled from some pdf ● Samples are most likely to be drawn from a region near a mode ● We can use the modes to cluster the data ● Clustering has many applications: filtering, segmentation, tracking, classification, and compression.
Why do we care about modes?
Why use mean shift for clustering? ● K-means needs to know how many clusters to use. Clusters data into voronoi cells. ● Histograms require bin size and number of bins ● Mixture models require information about pdf structure
Intuition ● We have a set of data that represents discrete samples of a distribution ● Locally we can estimate the density of the distribution with a function ● Compute the gradient of this estimation function ● Use gradient ascent to find the peak of the distribution
How does it work? ● We estimate the density using: ● Where h (bandwidth) is the region around x where we are trying to estimate the density and k is some kernel function ● Instead of using the gradient of f, we use the mean shift vector:
How to find a mode? 1. Start at any point 2. Compute mean shift 3. if mean shift is zero: possible mode found 4. else move to where mean shift is pointing go to 2 ● To find multiple modes we need to try all points that are more than h distance apart ● Prune modes by perturbing them and checking for convergence ● Combine modes that are close together. Take the higher one.
How to cluster using mean shift? ● Every point in the data set will converge to some mode using mean shift ● We cluster points together if they converge to the same mode
Mean Shift Filtering ● Cluster using intensity and position. Then change the intensity to match the cluster.
Mean Shift Segmentation ● Cluster using intensity and spatial information. Each cluster represents a segment of the image.
Discussion ● Need to select bandwidth and kernel function ● Gaussian kernel performs better, but takes longer to converge ● Kernel density estimation does not scale well with the dimension of the space.
Questions

Download Presentation

Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

HOLY SHIFT! Linda Zheng Roadmap You are here My Shift Introduction Shift AST Experience

HOLY SHIFT! Linda Zheng Roadmap You are here My Shift Introduction Shift AST Experience Overview with Shift What is the Shift Project? Abstract Syntax Tree (AST) Binary Expression Identifier Identifier PLUS a + b a b

632 views • 28 slides

1 2 nd Shift Associates 2 nd Shift Associates 3 rd Shift Associates 3 rd Shift Associates 2

Office Staff Office Staff 1 st Shift Associates 1 st Shift Associates 1 2 nd Shift Associates 2 nd Shift Associates 3 rd Shift Associates 3 rd Shift Associates 2 This is the whole reason every associate here goes home at the end of the day in

338 views • 10 slides

Guiding Principles System Shift: Racial Equity System Shift: A Connected & System Shift: A

Brenda Martinek, Chief of Student Support Services Dan Jung , Chief of Operations Curtis Wilson , Principal Benson Tech David Roy , Senior Director of Communications Guiding Principles System Shift: Racial Equity System Shift: A Connected &

363 views • 10 slides

Secondary Motivation #GOInnovate17 A Shift in Thinking A Shift in Perspective What do I need

Secondary Motivation #GOInnovate17 A Shift in Thinking A Shift in Perspective What do I need to know about the BRAIN? 4 Syncing Up 5 Image Source: Amazon.com Self-Regulation 6 Whats the deal with these hemispheres? Left Integrated

620 views • 15 slides

THE LARGEST SHIFT IN THE WAY WE LIVE - EVER AT THE SAME TIME BIRTHRATES Demographic shift. It is a

THE LARGEST SHIFT IN THE WAY WE LIVE - EVER AT THE SAME TIME BIRTHRATES Demographic shift. It is a COLLAPSE WORLDWIDE remarkable thing. We are experiencing the most rapid change in the nature of human habitat ever seen. Within the next five

368 views • 33 slides

Shift Work and the Impact on Wellbeing Helen Lawson Objectives Shift work in context &

Shift Work and the Impact on Wellbeing Helen Lawson Objectives Shift work in context & key guidance (UK) Known negative impacts associated with shift work How to support shift workers using a wellbeing framework The

379 views • 26 slides

VHDL Modeling for Synthesis Hierarchical Design Textbook Section 4.8: Add and Shift Multiplier

VHDL Modeling for Synthesis Hierarchical Design Textbook Section 4.8: Add and Shift Multiplier Add and shift binary multiplication Shift & add Shift & add Shift & add System Example: 8x8 multiplier Multiplicand Multiplier

297 views • 28 slides

Why the need for a paradigm shift? After many years of research on efficiency, ETL, highly

The road to highlights is paved with good intentions: envisioning a paradigm shift in OLAP modeling Panos Vassiliadis Patrick Marcel University of Ioannina, Hellas University of Tours, France Why the need for a paradigm shift? After many

272 views • 14 slides

Making the Shift K-9 Communicating Student Learning Why Shift? Click to add text Where are

Making the Shift K-9 Communicating Student Learning Why Shift? Click to add text Where are they going to next? Jennifer % Teresa % 55/60 = 92% 23/60 = 38% Formative assessment pretest 24/25 96% 16/25 64% Oral presentation

700 views • 41 slides

Power Shift Driving Change Report Webinar November 2017 Power Shift Helping consumers

Power Shift Driving Change Report Webinar November 2017 Power Shift Helping consumers manage their energy usage Power Shift Objectives Improve our evidence-based understanding of what really works in supporting vulnerable consumers to

884 views • 54 slides

Power Shift Multiple Impacts Report Webinar December 2017 Power Shift Helping consumers

Power Shift Multiple Impacts Report Webinar December 2017 Power Shift Helping consumers manage their energy usage Power Shift Objectives Improve our evidence-based understanding of what really works in supporting vulnerable consumers

718 views • 46 slides

Circular q-shift - Hypercube Using E-cube routing q-shift in a hypercube with p nodes:

Circular q-shift - Hypercube Using E-cube routing q-shift in a hypercube with p nodes: longest path has log p - (q) links, where (q) is the highest integer j, such that q is divisable by 2 j Pairwise communication using E- cube

303 views • 5 slides

1 Mean Shift Algorithm Graph-Theoretic Image Segmentation Mean Shift Algorithm 1. Choose a

6.891 Last time: Segmentation and Clustering (Ch. 14) Computer Vision and Applications Supervised->Unsupervised Category Learning needs segmentation Prof. Trevor. Darrell K-Means Mean Shift Graph cuts Lecture 15:

531 views • 19 slides

Modal Shift for an Environmental Lift A taxonomy of modal shift policy in Europe Johanna Takman 1

Foto Denys Bilytskyi, Mostphotos Modal Shift for an Environmental Lift A taxonomy of modal shift policy in Europe Johanna Takman 1 , Cecilia Gustafsson 1 , Inge Vierth 1 , Noor Sedehi Zadeh 1 , Anastasia Christodoulou 2 , Marta Gonzalez-Aregall 2

529 views • 21 slides

Operator Description Associtivity << bitwise left shift >> bitwise right

Operator Description Associtivity << bitwise left shift >> bitwise right shift (sign extend) [ ] indexing >>> bitwise right shift (zero pad) member selection . <, >, <=, numeric

82 views • 4 slides

The Shift Method Experience YOUR LIFE TRANSFORMATION BY: TUESDAY ROBBINS

The Shift Method Experience YOUR LIFE TRANSFORMATION BY: TUESDAY ROBBINS www.tuesdayrobbins.com How will this experience help you? The Shift Method will help you shift out of bad habits, limiting beliefs, and emotions that dont serve

483 views • 13 slides

ICS 667 Advanced HCI Design Methods 08. Intro to Evaluation Analytic Evaluation Dan Suthers

ICS 667 Advanced HCI Design Methods 08. Intro to Evaluation Analytic Evaluation Dan Suthers Spring 2005 Outline Introduction to Evaluation Types of Evaluation etc. Usability Specifications Analytic Methods Metrics and

519 views • 21 slides

Spoken Document Retrieval and Browsing Ciprian Chelba OpenFst Library C++ template library

Spoken Document Retrieval and Browsing Ciprian Chelba OpenFst Library C++ template library for constructing, combining, optimizing, and searching weighted finite-states transducers (FSTs) Goals: Comprehensive, flexible, efficient and

892 views • 36 slides

The rms-flux relation In Black Hole Binaries Credit: ESA Lucy Heil (Leicester) With: Simon

The rms-flux relation In Black Hole Binaries Credit: ESA Lucy Heil (Leicester) With: Simon Vaughan & Phil Uttley (Leicester) (Southampton) The RMS-flux relation V4641 First identified in Uttley & McHardy (2001) XRB Cyg X-1

516 views • 14 slides

On Separators in Temporal Graphs Hendrik Molter Algorithmics and Computational Complexity, TU

On Separators in Temporal Graphs Hendrik Molter Algorithmics and Computational Complexity, TU Berlin, Germany Algorithmic Aspects of Temporal Graphs, Satellite Workshop of ICALP 2018, Prague Based on joint work with Till Fluschnik, Rolf

1.99k views • 174 slides

In the previous parts of the course, weve seen a number of different estimators used in

In the previous parts of the course, weve seen a number of different estimators used in volumetric light transport: bidirectional path tracing, volumetric photon mapping, the beam radiance estimate, photon beams etc. In this part of

901 views • 46 slides

Model selection for fast density estimation orfi 1 L aszl o (Laci) Gy 1 Department of

Model selection for fast density estimation orfi 1 L aszl o (Laci) Gy 1 Department of Computer Science and Information Theory Budapest University of Technology and Economics Budapest, Hungary July 17, 2008 e-mail: gyorfi@szit.bme.hu

951 views • 68 slides

Instance-based Learning CE-717: Machine Learning Sharif University of Technology M. Soleymani

Instance-based Learning CE-717: Machine Learning Sharif University of Technology M. Soleymani Fall 2018 Outline } Non-parametric approach } Unsupervised: Non-parametric density estimation } Parzen Windows } Kn-Nearest Neighbor Density

863 views • 52 slides

Density Estimation Parametric techniques Maximum Likelihood Maximum A Posteriori

1 Density Estimation Parametric techniques Maximum Likelihood Maximum A Posteriori Bayesian Inference G aussian M ixture M odels (GMM) EM-Algorithm Non-parametric techniques Histogram Parzen Windows

496 views • 11 slides