Computer Vision Module 5

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

Applications : Photo album – Face detection – Face recognition – Eigen

faces – Active appearance and 3D shape models of faces. Application:


Surveillance – foreground-background separation – particle filters –
Chamfer matching, tracking, and occlusion – combining views from
multiple cameras – human gait analysis & Vehicular detection in tolls.
PHOTO ALBUMS
▪ The goal of computer vision is to understand the story unfolding in a picture. As humans, this is quite
simple. But for computers, the task is extremely difficult.

▪ Images are everywhere! Whether it be personal photo albums on your smartphone, public photos on
Facebook, or videos on YouTube, we now have more images than ever.

▪ The need for methods to analyse, categorize, and quantify the contents of these images arises.

▪ For example, have you recently tagged a photo of yourself or a friend on Facebook lately? How does
Facebook seem to “know” where the faces are in an image? Facebook has implemented facial
recognition algorithms into their website, meaning that they cannot only find faces in an image, they can
also identify whose face it is as well!

▪ Facial recognition is an application of computer vision in the real world.


DEFINITION
• A digital photo album is a collection of digital images, that is stored on an
electronic device, and can easily be manipulated.
• It may or may not be portable.
▪ A digital photo album is a software interface that allows users to upload, store
and often manipulate digital images.
▪ Other features may include arranging and labelling, as well as specific viewing
options.
▪ Digital photo albums can be run as standalone applications, or as part of an
operating system.
▪ They might also come paired with camera packages.
▪ Many popular websites and social media platforms offer digital photo albums, as
either a flagship or secondary feature.
HISTORY
▪ Digital photography can be traced back to the 1960s, with the development of digital imaging and
the charge-coupled device (CCD).

▪ Solid state and SLR (single-lens reflex) cameras, combined with personal computing’s rise, led to
the digital photo album becoming a staple among professional and amateur photographers alike.

▪ In 1988, image editing software called Digital Darkroom was released for Macintosh and was one
of the first digital photograph computer programs.

▪ All major computer operating systems have included some form of digital photo album in their
platforms, with many automatically enabled and functional immediately upon installation
ROLE OF COMPUTER VISION
▪ There are many software programs available to manipulate and manage images and videos in folders or
albums. The functionalities provided by the concepts of Computer Vision is explained below.

▪ These programs generally allow for sorting and ordering of different images, tagging the images, and
viewing them in slideshows or printing them. In terms of videos, shortening, addition of visual effects,
merging different clips, etc. are a few facilities that are provided in these programs.

▪ These programs commonly allow the user to perform basic edits such as cropping, red-eye removal, and
some basic "one touch" enhancements for colour and lighting.

▪ Automatic syncing all of the photos and videos from your smartphone/system to a secure location in the
cloud is provided as well.
Face detection
• Over the past few years, face recognition was considered one of the
most promising fields in image analysis.
• Before face recognition can be applied, the size and the shape of the
face is to be detected.
• This is the core principle of face detection.
• There are various types of face detection techniques depending upon
where it is used.
• Computer vision uses the applications of face detection extensively
and in many day-to-day lives.
Types of face detection
Feature based
• The feature-based method is to locate faces by extracting
structural features of the face.
• They can be used to extract distinct image features like eyes,
nose and so on.
• By doing so, it is then used to differentiate between facial and
non-facial regions.
• The idea is to overcome the limits of our instinctive knowledge
of faces.
• They may detect various aspects of the image like:
• Edges
• Corners
• Blobs
• Ridges
Template based
• Template Matching method uses pre-defined or parameterized
face templates to locate or detect the faces by the correlation
between the templates and input images.
• A face model can be built by feature extraction and then we
need to apply template matching to get the desired output.
• This approach is simple to implement, but it is inadequate for
face detection.
Template matching
Appearance based
• The appearance-based method depends on a set of delegate
training face images to find out face models.
• It scans over small overlapping rectangular patches of the
image searching for matches.
• Most appearance based classifiers rely mostly on training
classifiers using sets of labelled face and non-face patches.
The image is viewed in multiple overlapping rectangular boxes
and then each part gets analyzed separately.
Appearance based
• There are many methods used in appearance based classification. Let
us take a look at some of them.
• Clustering :-
• It is the grouping together of data into similar groups. Doing this,
we find out the group to which a particular image or a part of the
image may belong to.
• We can find out various features and parts of the image doing the
same.
• It is very effective but needs a lot of training data.
Appearance based
• Neural-Networks:-
• Instead of clustering, we use a neural network that has numerous inputs and
learns from them and gives us the output.
• Many detection problems like object detection, face detection, emotion detection,
and face recognition, etc. have been faced successfully by Neural Networks.

• Support Vector Machine:-


• Support Vector Machines are linear classifiers that maximize the margin
between the decision hyperplane and the examples in the training set.
• They search for a series of maximum margins separating planes in
feature space between different classes.
Face Recognition
• Face Recognition is a recognition technique used to detect faces of individuals
whose images saved in the data set. Despite the point that other methods of
identification can be more accurate, face recognition has always remained a
significant focus of research because of its non-meddling nature and because it is
people’s facile method of personal identification.
Face Recognition
Face Recognition
• Face recognition systems are usually composed of the following building blocks:

1) Face detection. A face detector finds the position of the faces in an image and (if any)

returns the coordinates a bounding box for each one of them. This is illustrated in Figure

2) Face alignment. The goal of face alignment is to scale and crop face images in the

same way using a set of reference points located at fixed locations in the image.

This process typically requires finding a set of facial landmarks using a landmark detector

and, in the case of a simple 2D alignment.


Face Recognition
• 3) Face representation. At the face representation stage,the pixel values of
a face image are transformed into a compact and discriminative feature
vector, also known as a template. Ideally, all the faces of a same subject
should map to similar feature vectors.
4) Face matching. In the face matching building block,two templates are
compared to produce a similarity score that indicates the likelihood that
they belong to the subject.

You might also like