Week 11 - Motion & Tracking Flashcards
(29 cards)
What is Tracking in Computer Vision?
Tracking involves following a specific object through a video
What are some issues with pixel-level motion tracking solutions?
- They do not make predictions for new locations
- They do not estimate velocities of objects/targets in the image
- They do not handle uncertainty of position, or challenges like occluding objects
What are Flow Fields used for?
It is used in tracking, and can be used to infer motion, stabilise images, stabilise a device or help with frame interpolation
What are the two types of approaches for Motion Detection?
- Motion Difference
- Background Subtraction
How does Motion Difference work in Motion Detection?
- Take two images from a sequence
- Compute the change in brightness at each pixel in the image
- Threshold and filter for noise
How does Background Subtraction work in Motion Detection?
- Capture an image of the background
- Use the difference between the current frame and the background to find moving objects
What is the best way to take background images for Background Subtraction models in Motion Detection?
Take several at different points in time, so that the average of a given pixel’s value over some time period is better
How do you work out the average value for a background pixel in Motion Detection?
Averaging often becomes a Gaussian model, where you compute deviation as well as just mean:
- Given a new pixel value and a Gaussian model, you can estimate the pixel value
What is an example algorithm that uses confidence to predict the movement of an object in Tracking, and how does it work?
Kalman Filter
- Handles certainty of measurements and prediction
- Has a motion model
- Assumes we have a unimodal Gaussian i.e. can only predict one location
What is a short description of a Particle Filter?
A particle filter uses a set of samples to approximate the target probability density.
How does a Particle Filter work?
A Particle Filter treats each sample as a particle, which is a hypothesis with some probability of being correct
What is the plain English description of the theory behind a Particle Filter?
- Start with many guesses about where the object could be
- Predict where each guess thinks the object could be
- Compare those guesses to what you actually see in the new image i.e. the measurement
- Update the scores of each guess based on how well they match the real observation
- Pick new guesses based on the best scores, and repeat
What is the basic algorithm for CONDENSATION in Tracking?
- Select a particle
- Project forward in time
- Add noise
- Compute weight from measurement
What are some features of CONDENSATION in Tracking?
- The hypotheses can have any shape probability distribution
- The observations are used to update the probabilities of the hypotheses via an appearance model
- Many hypotheses are considered at once, which increases chances of finding the target
- Works in substantial clutter at or close to video frame rate
- Can use different numbers of particles as a trade-off between running very fast and having higher accuracy and greater robustness
What is an Appearance model in CONDENSATION?
An appearance model is like a predicted silhouette of what the object will look like. You come up with a measurement which is something like this example: 1 if over a red pixel patch, tends to 0 as you drift from this.
What is Contour Tracking?
Contour Tracking establishes a state that contains a set of points representing a contour, so each particle represents a possible contour
This state has higher dimensions, which are called the parameters of the contour.
The mean state is usually drawn to summarise results
What is a mixed-state CONDENSATION model?
It saves multiple transition probabilities of an object switching behaviour, and then builds multiple models using this data into a single CONDENSATION tracker.
What is the step-by-step process for a Mixed-State CONDENSATION tracker?
- Select a particle, which includes a gamma(t)
- Randomly generate gamma(t + 1) in line with the transition probability table
- Project particle forward with motion model associated with gamma(t + 1)
- Add noise
- Compute pi(t)
What happens to the particles in a Mixed-State CONDENSATION tracker?
- Mixed-state CONDENSATION produces a mix of particles, depending on the transition probabilities.
- Different particles contain and follow different motion models
- Clouds of particles can form
What is an issue with using transition tables for transition probability predictions in tracking?
It was hard-coded, and as such, is difficult to do and is inflexible
How does the number of particles affect the performance of particle filter methods?
Lower number:
- Fast, but coarse
- May have unsampled areas
Higher Number:
- Slower
- More potential for representing state space
What affects the use of CONDENSATION models/trackers?
The Curse of Dimesionality
What is MCMC tracking, and what are some key properties?
Markov Chain Monte Carlo:
- Allows us to explore the space of possible tracker/location states more efficiently
- Don’t apply the motion model to everything inside the state
- Improve a particle by selecting one target inside it, and moving only that and seeing if its better or worse than before
- Only add improved samples to the new particle set
- Effectively a random walk looking for improved target configurations
How does the MCMC tracking work?
- Generate a new set of samples using the ‘Metropolis-Hastings’ algorithm:
– Propose a new state
– Evaluate the acceptance ratio
– If the acceptance ratio is greater than 1, then accept the state and update the target we considered
– Otherwise, accept it if the random variation is high enough
– Otherwise, keep the current state and reject the proposed state
