Computer Vision: Detection, Segmentation, and Applications

A set of flashcards covering classification, localization, detection, segmentation algorithms, high-level processing, and various real-world applications of computer vision.

Detection and segmentation

Fundamental tasks in computer vision focusing on identifying objects or regions of interest within an image.

Single Object Tasks

Tasks focusing on analyzing or delineating individual objects within an image, including classification and localization.

Classification

The task of determining the category or class to which a single object belongs to provide insights into its identity.

KNN (K-Nearest Neighbour)

An algorithm that may be used for supervised classification in computer vision.

K-means clustering

An algorithm that can be used for unsupervised classification in computer vision.

Object Localization

Involves identifying the class label of an object as well as a bounding box to show where the object is located.

Classification + Localization

A task that involves both classifying an object and precisely localizing it by predicting bounding boxes that tightly enclose it.

Multiple Object Tasks

Tasks that deal with scenarios where an image contains multiple instances of objects or different object classes to identify and distinguish between them.

Object Detection

Focuses on identifying and locating multiple objects of interest by analyzing the entire image, drawing bounding boxes, and assigning class labels.

R-CNN

Region-Based Convolutional Neural Network; an algorithm used for object detection.

R-FCN

Region-based Fully Convolutional Network; an algorithm used for object detection.

YOLO

You Only Look Once; a popular algorithm used for object detection.

SSD

Single Shot Detector; an algorithm used for object detection.

Image segmentation

The process of creating a mask around similar characteristic pixels and identifying their class to understand an image at a granular level.

Edge detection

A technique used in image segmentation that works by detecting discontinuities in brightness.

Semantic Segmentation

A type of segmentation that classifies pixels belonging to a particular class without differentiating between individual objects of the same class.

Instance Segmentation

A type of segmentation that classifies pixels belonging to a particular instance, differentiating all objects even if they belong to the same class.

High-Level Processing

The final stage of computer vision involving interpreting and extracting meaningful information to recognize objects, understand scenes, and analyze context.

Facial recognition

A computer vision application used by social media platforms like Facebook to detect and tag users.

Optical character recognition (OCR)

An application used to extract printed or handwritten text from visual data such as images, invoices, or articles.

Machine inspection

A process that detects a machine's defects, functional flaws, and irregularities in manufactured products.

3D model building

Constructing 3D computer models from existing objects for use in Robotics, Autonomous driving, and AR/VR.

Surveillance

The use of live CCTV footage to identify suspicious behavior, detect dangerous objects, and prevent crimes.

Fingerprint recognition and biometrics

Applications that detect fingerprints and biometrics to validate a user's identity.