Questions tagged [computer-vision]

Questions about algorithms for processing and analyzing images and videos.

112 questions with no upvoted or accepted answers
Filter by
Sorted by
Tagged with
7
votes
0answers
2k views

Detecting the damaged regions in cars

Detecting the regions where a car has been damaged and the extent to which it has been damaged is a very interesting problem. It has potential applications in automatic auto insurance claims. ...
6
votes
1answer
296 views

Why is Scale Space Needed to Detect Scale Invariant Features?

I have been doing research on the SIFT (scale invariant feature transform) algorithm, and while I have repeatedly read that in order to detect image features that do not vary with respect to scales of ...
3
votes
1answer
35 views

Matching superimposed image

We are given two grayscale images, one of which contains a large, mostly contiguous patch from the other one. The patch can be altered with noise, its levels may be stretched, etc. Here's an example ...
3
votes
0answers
17 views

Pattern detection by stacking convolutions

CNN use convolutions to perform many tasks. In some architecture, the first layer(s) are basic filters such as Gabor's. Many CNN work by stacking a number of convolution layers to build detectors ...
3
votes
0answers
43 views

Partial derivatives of bilinear Sampling

I'm going over the paper Spatial Transformer Networks (link). On section 3.3 they introduce bilinear sampling kernel (Eq. 5) $$ V_{i}^{c}=\sum_{n}^{H} \sum_{m}^{W} U_{n m}^{c} \max \left(0,1-\left|x_{...
3
votes
0answers
27 views

Using colors in feature detection

Is there any feature detector or extractor that uses color information? Detectors like SURF and SIFT and many others work with grayscale. How do you think, may be transform can be sued to generate a ...
3
votes
0answers
63 views

Calculating depth mask from different lighting

I have a object which is static, the camera is static and light source is moving. How can the depth mask be calculated ? Concept is to use - calculate height from shadow length Lets imagine a have ...
3
votes
0answers
20 views

impact of the single pixel in disparity

Disparity essentially means the distance of a corresponding pixel in another image ( in pixels). This is clear and understood. When I looked after how the depth is related to disparity, then I found ...
3
votes
1answer
216 views

Image registration using gradient descent

I have a target image $f(x,y)$ (where $x \in [0, 250]$ and $y \in [0,300]$), and a source image $g(x,y)$ I want to align $g$ to $f$ using the transformation : $$\Psi(x,y;t_x, t_y, \theta) = \begin{...
3
votes
0answers
102 views

How can color spaces be compared?

I am currently working on image classification methods. One group of preprocessing methods are color space transformations. While there are "traditional" color spaces like RGB (the raw data), HSV, LUV,...
3
votes
1answer
639 views

HandyAR finger tip detection and tracking

I am reading this paper Handy AR. As suggested I am adding certain quotes from paper to get basic idea of what is it about : Intro In this paper, we present and evaluate a method to use a user’...
2
votes
0answers
24 views

L1 loss different scale for small or large boxes [object detection]

In this paper[1] there is a section where it is mentioned that “The most commonly used L1 loss will have different scales for small and large objects even if their relative errors are similar”. I don’...
2
votes
0answers
17 views

Contrast detection implementation for blur detection in microscopic image

I am currently doing a project where I have to detect if the the microscopic images given are blurry. I tried canny edge detection but they seem to perform not so good in detecting blur. I realised ...
2
votes
0answers
48 views

Do image artefacts reveal the original image orientation?

An image is taken and then rotated by an arbitrary angle and cropped. Let us assume that the original image is rather meaningless and is not aligned in any way. Given the rotated image, is it ...
2
votes
0answers
32 views

Removing fish-eye effect

The DNG specification (page 87) describes a simple algorithm for converting a "fish-eye" photo into a regular photo (just several lines of basic geometry). The transformation is configured using six ...
2
votes
1answer
129 views

Do white box methods for feature extraction exist for facial recognition?

Part of something I'm working on (or rather begin working on) involves white box approaches to facial recognition. I previously worked on something for character recognition using scene segmentation ...
2
votes
0answers
23 views

How to find the matched SIFT features that are spatially consistent?

I have extracted DenseSIFT from the query and database image and quantized by kmeans. The challenge is to find those SIFT features that quantized to the same visual ...
2
votes
0answers
55 views

How the SIFT results vary by changing some parameters

I'm studying the SIFT algorithm and I have some problem understanding its operation, in particular I don't understand how the results vary according to the parameters: number of octaves, number of ...
2
votes
0answers
246 views

Geometric interpretations of midpoint algorithm, homogeneous linear least squares and nonlinear least squares method in 3D reconstruction?

In "Multiple View Geometry in Computer Vision" Chapter 12. Structure Computation, page 310-313, triangulation is used for point 3D reconstruction. There are three methods mentioned: Midpoint method ...
2
votes
0answers
495 views

Class activation mapping using inception-v3 network

I am using the technique of this paper: https://arxiv.org/pdf/1604.07953.pdf to localize food objects. The difference is that I am using the next generation of google-net, inception-v3 network. For ...
2
votes
0answers
107 views

Image properties given by similitude moments

What image properties are given by similitude moments? These are image moments of the form shown below: $$\eta_{pq}=\frac{\sum_x\sum_y(x-\mu_x)^p(y-\mu_y)^qI(x,y)}{(\sum_x\sum_yI(x,y))^{\frac{p+q}{2}+...
2
votes
0answers
683 views

Relative pose estimation using epipolar geometry: Wrong translation estimate

I am trying to use epipolar geometry concepts to estimate the relative pose of cameras between two images. I am currently using a simulator to get my camera images, so the distortion is zero and I ...
2
votes
0answers
48 views

Registering 3D-NIR image to thermal image and vice versa

In the past I have thought a bit about how to register a NIR-image and a thermal image and noticed that this is not trivial - one statement was that if I had the depth information for each pixel, the ...
2
votes
0answers
97 views

Learning Image Processing/Computer Vision

What do I need to learn dense correspondence? I am a newbie in Image processing/CV. Please help me by suggesting a kind of syllabus or show me a way to learn about Dense Matching. What are the things ...
2
votes
0answers
272 views

How to select a good window size in the Running Gaussian Average method for Selective Background Updating?

I have been studying the Selective Background Updating in Running Gaussian Average Method, as discussed in the Wikipedia (the Variant method). The details of the algorithm, as I understand, is as ...
1
vote
0answers
9 views

Methods to interpolate between 2 topologically identical 3D meshes

I have 2 3D surface meshes. These meshes have vertex-correspondence and have the same topology (same edges and triangles connecting the vertices). However, the vertex positions (3d coordinates) are ...
1
vote
0answers
11 views

Requesting help for monitoring data drift in instance segmentation model

happy to be here. I need some guidance. I am writing a project on data drift. I have trained an instance segmentation model on the Cityscapes Dataset, to detect pedestrians. The model I trained was on ...
1
vote
0answers
17 views

Disparity calculations and camera centers

I'm learning about the disparity calculations and want to calculate the distance for a known pixel. I have calibrated and rectified images already. Now I just want to calculate Z for some pixel x,y. ...
1
vote
0answers
20 views

How to calculate image gradient with respect to homogeneous coordinate vector?

I want to use the direct method of computing the homography between two images to estimate the plane parameters (normal and depth [= distance from camera]) of a 3D plane. If we have two images of the ...
1
vote
1answer
41 views

Is the Jacobian matrix calculation process correctly on this energy function?

When I implemented the paper Dynamic Fusion published in 2015 CVPR, I came into the trouble with how to solve the energy function: $$E(W_t, V, D_t, \mathcal{E}) = \text{Data}(W_t, V, D_t) + λ\text{Reg}...
1
vote
0answers
94 views

What is the best way to compare two images for similarity if the brightness of the second image was changed?

I calculate similarity percent between two images. Another image is the same with changed brightness. I've already tried comparing by pixel (euclidean distance for grayscale and 0/1 values), comparing ...
1
vote
0answers
15 views

How to correct distortion from stitching dual camera to estimate homography

I am trying to distort the image obtained from two cameras that generate a stitching image. Unfortunately, I do not have access to the cameras and I would like to know what could be the best way to ...
1
vote
0answers
13 views

Drawing a planar intersection from stereo camera

I wanted to recreate the results of this paper, where the authors draw the boundary between two parts of a building from stereo images. The authors mention that the epipole and the intersection of the ...
1
vote
0answers
30 views

Would a five color display have higher image quality than an RGB display?

We have five types of photo receptors in the eyes but computer color systems just use the three red, green and blue. Three could be motivated if rod and melanopsin sensations are just linear ...
1
vote
0answers
91 views

trouble recovering rotation and translation from essential matrix

I am having trouble recovering rotation and translation from an essential matrix. I am constructing this matrix using the following equation: \begin{equation} E = R \left[t\right]_x \end{equation} ...
1
vote
0answers
19 views

How are epipolar lines useful in computer vision?

Why do use epipolar lines and what are it's benefits in computer vision ?I am unable to find the uses of epiploar lines and how they are helpful although i am aware of epiploar geometry concepts. I ...
1
vote
0answers
35 views

3D Reconstruction from multiview calibrated camera

I have a calibrated camera whose intrinsics were calculated prior. Suppose I have multi images with the same camera. Using the first two views.I have the 1) initial reconstruction using 8 point ...
1
vote
0answers
30 views

How can I compare two faces using facial landmarks?

I'm using first_order_model for an expression transfer between an video and a static image. Part of the algorithm makes a facial comparison to get the frame in the video where the face is best ...
1
vote
0answers
21 views

What is the best algorithm for detecting the texture of a person's clothes in real-time for a mobile robot?

I am working on the human following robot. the robot extracts a feature of clothes's texture and distinguishes between the people that they stand up in front of the robot. So, the robot needs that it ...
1
vote
0answers
39 views

Image/Video deblurring

I am currently working on an algorithm that will help to improve shaken blurry images. I know there are plenty of them, such as deconvolution, that is pretty useful though I don't really know how to "...
1
vote
0answers
61 views

Why HOG angular bins have values between 0 and 180 degrees?

I have learned that HOG angular bins have values from 0 to 180 since the algorithm uses unsigned gradients. I am failing to understand why. As far as I know gradient orientation is calculated as <...
1
vote
0answers
70 views

RGBD alignment without explicit transformation between RGB and depth

I have a set of RGB-D scans of the same scene, corresponding camera poses, intrinsics and extrinsics for both color and depth cameras. RGB's resolution is 1290x960, depth's is 640x480. Depth is ...
1
vote
0answers
12 views

Descriptors in key-features during mapping

This question is related to this question. In SLAM-type applications the feature detectors like SIFT are used for finding keypoints in the images to be matched. The candidate keypoints are matched by ...
1
vote
0answers
61 views

Where is the epipole if one camera center is not in view of the other?

In the book multiple view geometry, the epipole is defined as follows: The epipole is the point of intersection of the line joining the camera centres (the baseline) with the image plane. ...
1
vote
0answers
15 views

Does the illumination color (red/blue/green) has any effect over the disparity map generated from a stereo image?

I have a working model that creates a disparity map using stereoscopic image processing using opencv. I am thinking of improving the disparity map by using illumination and stuck with different ...
1
vote
0answers
63 views

What is the common way to approximate contour with quad?

I'm trying to find quads on stream from my webcam. cv2.HoughLinesP sometimes returns more lines then needed. How can I approximate them to contour of exactly 4 points that fits returned result best? I'...
1
vote
0answers
59 views

Why is sparsity useful in dictionary learning?

What is the particular reason for which we use sparse matrix X in dictionary learning equation Y=DX where Y is the signal matrix, D is dictionary and X is the coefficient matrix?
1
vote
0answers
34 views

Finding the linear mapping between homogeneous coordinates of affine camera

If I have an affine camera with a projection relationship governed by: \begin{equation} \begin{bmatrix} x & y \end{bmatrix}^T = A \begin{bmatrix} X & Y & Z \end{bmatrix}^T + b \end{...
1
vote
0answers
26 views

3D registration for reconstructed surface

I use Meshroom (based on the framework AliceVision) to reconstruct a scene based on photographs (e.g. a bottle on a table in a room). The software captures the bottle, but also the table and the whole ...
1
vote
0answers
90 views

Algorithm to determine paths that define shape of connected rectangles

Say you have a bunch of rectangles like these: They get organized into 3 distinct final shapes like this: As you can see, there are a few "complete contours", as in here: The way you figure out the ...