Computer Vision III: Detection, Tracking, Segmentation at TU München

Flashcards and summaries for Computer Vision III: Detection, Tracking, Segmentation at the TU München

Arrow Arrow

It’s completely free

studysmarter schule studium
d

4.5 /5

studysmarter schule studium
d

4.8 /5

studysmarter schule studium
d

4.5 /5

studysmarter schule studium
d

4.8 /5

Study with flashcards and summaries for the course Computer Vision III: Detection, Tracking, Segmentation at the TU München

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Briefly describe Viola-Jones Detector

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Briefly describe procedure of Histogram of Oriented Gradients

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Name Three Classical Object Detection Methods

This was only a preview of our StudySmarter flashcards.
Flascard Icon Flascard Icon

Millions of flashcards created by students

Flascard Icon Flascard Icon

Create your own flashcards as quick as possible

Flascard Icon Flascard Icon

Learning-Assistant with spaced repetition algorithm

Sign up for free!

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Briefly describe the algorithm of Non-Max Suppression 

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Name several one-stage detectors

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Name several two-stage detectors

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Which datasets are used to evaluate object detection methods?

This was only a preview of our StudySmarter flashcards.
Flascard Icon Flascard Icon

Millions of flashcards created by students

Flascard Icon Flascard Icon

Create your own flashcards as quick as possible

Flascard Icon Flascard Icon

Learning-Assistant with spaced repetition algorithm

Sign up for free!

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Describe the training scheme for R-CNN

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

What is the main contribution of the SPP-Net?

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

How does the Fast R-CNN work?

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

How does a Region Proposal Network work?

This was only a preview of our StudySmarter flashcards.
Flascard Icon Flascard Icon

Millions of flashcards created by students

Flascard Icon Flascard Icon

Create your own flashcards as quick as possible

Flascard Icon Flascard Icon

Learning-Assistant with spaced repetition algorithm

Sign up for free!

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

What is YOLO and how does it work?

Your peers in the course Computer Vision III: Detection, Tracking, Segmentation at the TU München create and share summaries, flashcards, study plans and other learning materials with the intelligent StudySmarter learning app.

Get started now!

Flashcard Flashcard

Exemplary flashcards for Computer Vision III: Detection, Tracking, Segmentation at the TU München on StudySmarter:

Computer Vision III: Detection, Tracking, Segmentation

Briefly describe Viola-Jones Detector

- learn multiple weak learners on Haar-like features

- combine result of weak learners to form strong final decision

- use AdaBoost to find best combination of weak learners

Computer Vision III: Detection, Tracking, Segmentation

Briefly describe procedure of Histogram of Oriented Gradients

- choose images that contain object you want to detect

- choose images that do NOT contain said object

- compute HOG for each image

- train SVM on HOG to separate images with object and images without object

Computer Vision III: Detection, Tracking, Segmentation

Name Three Classical Object Detection Methods

- Template matching + sliding window

- Viola-Jones Detector

- Histogram of Oriented Gradients (HOG)

Computer Vision III: Detection, Tracking, Segmentation

Briefly describe the algorithm of Non-Max Suppression 

- Algo 

for box in AllBoxes:

  for otherBox in AllBoxes:

   if overlap(box, otherBox):

    keepBetterBox(score(box), score(otherBox))


- overlap is IoU

- score depends on task

- NMS used in many tasks


Computer Vision III: Detection, Tracking, Segmentation

Name several one-stage detectors

- YOLO

- Single-Shot Multiscale Detector (SSD)

- RetinaNet

- CornerNet

- CenterNet

- ExtremeNet

Computer Vision III: Detection, Tracking, Segmentation

Name several two-stage detectors

- R(egion)-CNN

- Fast R-CNN

- Faster R-CNN

- SPP-Net

- R-FCN

- FPN

Computer Vision III: Detection, Tracking, Segmentation

Which datasets are used to evaluate object detection methods?

- PASCAL VOC 2007-12

- ImageNet 2010-17

- COCO 2015-

- OpenImages 2018-

Computer Vision III: Detection, Tracking, Segmentation

Describe the training scheme for R-CNN

1. Pretrain CNN on ImageNet

2. Fine-tune CNN on classes that you are actually trying to predict

3. Train one SVM per class on image regions.

4. Train bounding box regressor (L2 loss)

Computer Vision III: Detection, Tracking, Segmentation

What is the main contribution of the SPP-Net?

Instead of passing every proposal region through a separate CNN, the entire image is passed through a single CNN

Computer Vision III: Detection, Tracking, Segmentation

How does the Fast R-CNN work?

- like SPP-Net it applies a CNN on entire image

- the proposals are then extracted from the feature map and warped using ROI pooling

- to warp the features, one applies a grid of fixed size on each proposal (backprop like max-pooling)

Computer Vision III: Detection, Tracking, Segmentation

How does a Region Proposal Network work?

- Per pixel in the feature map, you have 9 anchors (3 differently sized boxes at 3 different ratios)

- the proposals get refined via regression (L1 loss) and classified as object/non-object

Computer Vision III: Detection, Tracking, Segmentation

What is YOLO and how does it work?

- You Only Look Once is a one-stage detector

- there are no object proposals but detection and box regression is done directly from feature map

- for this, the image is divided into grid and multiple boxes are set in the grid (~100k proposals)

Sign up for free to see all flashcards and summaries for Computer Vision III: Detection, Tracking, Segmentation at the TU München

Singup Image Singup Image
Wave

Other courses from your degree program

For your degree program Computer Vision III: Detection, Tracking, Segmentation at the TU München there are already many courses on StudySmarter, waiting for you to join them. Get access to flashcards, summaries, and much more.

Back to TU München overview page

Computer Vision

Computer Training

computer vision

Computer Training

Computer Training

What is StudySmarter?

What is StudySmarter?

StudySmarter is an intelligent learning tool for students. With StudySmarter you can easily and efficiently create flashcards, summaries, mind maps, study plans and more. Create your own flashcards e.g. for Computer Vision III: Detection, Tracking, Segmentation at the TU München or access thousands of learning materials created by your fellow students. Whether at your own university or at other universities. Hundreds of thousands of students use StudySmarter to efficiently prepare for their exams. Available on the Web, Android & iOS. It’s completely free.

Awards

Best EdTech Startup in Europe

Awards
Awards

EUROPEAN YOUTH AWARD IN SMART LEARNING

Awards
Awards

BEST EDTECH STARTUP IN GERMANY

Awards
Awards

Best EdTech Startup in Europe

Awards
Awards

EUROPEAN YOUTH AWARD IN SMART LEARNING

Awards
Awards

BEST EDTECH STARTUP IN GERMANY

Awards
X

StudySmarter - The study app for students

StudySmarter

4.5 Stars 1100 Rating
Start now!
X

Good grades at university? No problem with StudySmarter!

89% of StudySmarter users achieve better grades at university.

50 Mio Flashcards & Summaries
Create your own content with Smart Tools
Individual Learning-Plan

Learn with over 1 million users on StudySmarter.

Already registered? Just go to Login