CSCI3290
Computational Imaging and Vision
Camera geometry, compositing and matting, HDR and gradient-domain imaging, stereo, image registration and motion, sampling and aliasing, deblurring and content-aware resizing.
Your progress
0/ 33 sections marked complete
Next section: 1.1 Cameras, homographies and image warping
Completion marks are saved in this browser.
This course in the web
In the CSCI3290 notes
- Theorem cards
- 64
- Definitions
- 74
- Counterexamples
- 35
- Concept lenses
- 26
- Histories of ideas
- 4
- Proof X-rays
- 33
Sections with a history of an idea
- HDR estimation, tone mapping and image-based lighting1.3
- Patch quilting, optimal seams and texture transfer1.26
- Patch synthesis and finite exemplar models1.27
- Calibrating a camera from several plane views1.33
None of these is on the timeline yet.
No concept lens from this course is recorded in the web yet.
Counts cover the 33 public sections of this course and what is written in them, not recorded connections. 0 of these sections are indexed into the web in detail so far.
In the webTheorem cards 64 · Definitions 74 · Counterexamples 35 · Concept lenses 26 · Histories of ideas 4 · Proof X-rays 33
None of these is on the timeline yet.
No concept lens from this course is recorded in the web yet.
Counts cover the 33 public sections of this course and what is written in them, not recorded connections. 0 of these sections are indexed into the web in detail so far.
Chapter 1 · 19 sections
Geometry, layers and light
A connected mathematical imaging selection with explicit information limits and twelve solved exercises per language.
Theorems 36 · Definitions 53Counterexamples 21Lenses 17History 4Proof X-rays 19
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 1.1
Cameras, homographies and image warping
Derive perspective projection, identify when a homography is valid, and distinguish geometric mapping from reconstruction and sampling.
- 1.2
Compositing, matting and coherent mosaics
Derive premultiplied compositing, identify the information needed for matting, and separate alignment, exposure and seam decisions in a mosaic.
- 1.3
HDR estimation, tone mapping and image-based lighting
Separate sensor measurements from display appearance, derive exposure fusion under an explicit noise model, and use radiance maps for diffuse illumination.
- 1.20
Exposure compensation on an overlap graph
Estimate relative image gains and diagnose inconsistent overlaps.
- 1.21
Thin lenses, defocus and depth of field
Derive a geometric blur circle and depth interval, and distinguish focus from sampling and exposure.
- 1.22
Sensor noise and exposure averaging
Derive a shot-and-read-noise model and follow uncertainty through averaging and demosaicing.
- 1.23
White balance and color constancy
Estimate channel gains from a neutral reference and examine the assumptions behind automatic white balance.
- 1.24
Edge-aware filtering and non-photorealistic stylization
Analyze bilateral-type weighted averages, prove their finite-grid guarantees, and construct a transparent smoothing, quantization and line-extraction pipeline.
- 1.25
Multiscale pyramid blending
Construct and reconstruct a finite image pyramid, blend its bands, and examine its limits.
- 1.26
Patch quilting, optimal seams and texture transfer
Construct a finite two-dimensional quilt, prove the overlap-seam recurrence, and use a target correspondence map without confusing local costs with image quality.
- 1.27
Patch synthesis and finite exemplar models
Build an explicit nearest-patch objective, follow a finite synthesis algorithm completely, and distinguish local descent from texture realism.
- 1.28
Image morphing: correspondence, triangles and foldovers
Construct a piecewise affine morph, prove what vertex correspondence guarantees, and diagnose collapse, overlap and resampling limits.
- 1.29
Estimating homographies and robust mosaic alignment
Construct normalized linear estimators, prove their rank conditions, calibrate a restricted camera model, and condition robust-sampling probabilities correctly.
- 1.30
Features, matching and global panorama adjustment
Analyze patch identifiability and normalized matching, then solve an anchored global alignment model with cycle inconsistency.
- 1.31
Cylindrical panorama warping
Derive forward and inverse cylindrical maps, their domains, and the limits of rotational image alignment.
- 1.32
Rotation panoramas and the obstruction of parallax
Derive the depth-independent rotation homography, test its assumptions, and explain why translation across several depths breaks one global warp.
- 1.33
Calibrating a camera from several plane views
Recover unknown intrinsics from plane homographies with explicit rank, pose, cheirality and distortion assumptions.
- 1.34
Radial distortion correction and calibration diagnostics
Fit a restricted radial model, establish its invertible field, and use it for inverse image resampling.
- 1.35
Two-view geometry, estimation and gauge
Derive a normalized linear two-view estimator, identify calibrated motion constraints and explain ambiguity, degeneracy and lens-model boundaries.
Chapter 2 · 1 section
Inverse imaging
Model and regularize an image reconstruction problem.
Theorems 2 · Definitions 1Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 2.1
Blur, inversion and regularization
Model blur, lost information and noise amplification; derive and assess regularized reconstruction.
Chapter 3 · 1 section
Stereo geometry
Recover depth under a calibrated model and assess matching uncertainty.
Theorems 1 · Definitions 1Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 3.1
Stereo disparity and depth
Derive rectified disparity, depth and uncertainty while examining correspondence limits.
Chapter 4 · 1 section
Sensor sampling
Study interleaved sensor measurements and their information limits.
Theorems 2 · Definitions 2Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 4.1
Color-filter sampling and demosaicing
Model interleaved color sampling, interpolation, identifiability and noise.
Chapter 5 · 1 section
Gradient-domain imaging
Match local image differences with fixed boundary values.
Theorems 2 · Definitions 1Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 5.1
Gradient-domain image blending
Derive an anchored gradient-matching problem and assess incompatible guidance.
Chapter 6 · 1 section
Image registration
Estimate periodic translation and identify correspondence limits.
Theorems 1 · Definitions 1Counterexamples 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 6.1
Phase correlation and image translation
Derive periodic phase correlation and assess shift ambiguity.
Chapter 7 · 1 section
Image motion
Estimate local image motion and assess its identifiability.
Theorems 1 · Definitions 1Counterexamples 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 7.1
Optical flow and the aperture problem
Derive local optical-flow constraints, least squares and rank limitations.
Chapter 8 · 1 section
Illumination and shape
Infer local normals under calibrated illumination.
Theorems 2 · Definitions 1Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 8.1
Photometric stereo and surface normals
Recover surface normals using calibrated illumination and rank conditions.
Chapter 9 · 2 sections
Sampling and aliasing
Analyze sampling ambiguity and prefiltering with explicit models.
Theorems 7 · Definitions 4Counterexamples 2Lenses 1Proof X-rays 2
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 9.1
Sampling, aliasing and prefiltering
Derive sampled aliases, finite-family recovery and prefiltering.
- 9.2
Convolution, linear filters and frequency response
Define correlation and convolution for sampled images, show why linear shift-invariant filters are convolutions, exploit separability, and read a kernel's effect from its frequency response.
Chapter 10 · 1 section
Content-aware resizing
Select and analyze minimum-energy image seams.
Theorems 2 · Definitions 2Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 10.1
Minimum-energy seams for image resizing
Derive and trace a dynamic program for one vertical image seam, with explicit objective and boundary conditions.
Chapter 11 · 1 section
Integral images
Rectangle sums and local image statistics.
Theorems 2 · Definitions 1Counterexamples 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 11.1
Integral images and local rectangle statistics
Derive a four-corner query and local moments, including boundary and numerical assumptions.
Chapter 12 · 1 section
Intensity distributions
Discrete cumulative histograms and intensity maps.
Theorems 2 · Definitions 2Counterexamples 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 12.1
Discrete histogram equalization and matching
Construct monotone intensity maps from empirical cumulative histograms and analyze quantization and ties.
Chapter 13 · 1 section
Compositing and matting
Linear-light premultiplied layers, masking limits and an inverse problem.
Theorems 3 · Definitions 3Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 13.1
Premultiplied alpha, source-over and the matting inverse problem
Derive linear-light compositing, prove its algebra and distinguish matting from forward blending.
Chapter 14 · 1 section
Camera geometry extensions
Derive calibrated rays and plane-induced maps, including degeneracies.
Theorems 1 · Definitions 1Counterexamples 1Lenses 1Proof X-rays 1
Chapter outline →See this chapter on the web →Not yet indexed in the web
- 14.1
Pinhole cameras: rays, planar homographies and parallax
Derive calibrated camera projection, recover the ray behind a pixel, and prove when a plane gives an invertible image homography.