DTA. M8. Lecture

Vocabulary flashcards covering the fundamental terms and concepts introduced in the lecture on computer vision and its applications in insurance and other industries.

Computer Vision

A field of artificial intelligence that enables computers to derive meaningful information from digital images or videos and make decisions or recommendations based on that information.

Artificial Intelligence (AI)

The broader discipline that seeks to create systems capable of performing tasks that normally require human intelligence, such as reasoning, learning, and perception.

Deep Learning

A subset of machine learning that uses multi-layered neural networks to automatically learn representations and patterns from data.

Neural Network

A computational model inspired by the human brain, composed of interconnected nodes (neurons) that process data in layers to recognize patterns.

Convolutional Neural Network (CNN)

A type of deep neural network especially effective at processing grid-like data such as images, using convolutional layers to detect spatial hierarchies of features.

Supervised Learning

A machine-learning approach where models are trained on labeled data so they can learn to map inputs to desired outputs.

Training Data

A curated set of labeled examples used to teach a model how to perform a specific task, such as identifying damaged versus undamaged vehicles.

Metadata

Descriptive information attached to data (e.g., image tags like damage type, repair cost) that aids training and interpretation by AI systems.

Object Detection

A computer-vision task that identifies and localizes all relevant objects within an image or video, typically drawing bounding boxes around them.

Bounding Box

A rectangle drawn around an object in an image to indicate its position and size for tasks like object detection.

Image Classification

Assigning a label or category to an entire image based on its visual content.

Image Restoration

Techniques used to improve or reconstruct the quality of an image that has been degraded or corrupted.

Accuracy Rate

The percentage of predictions a model gets correct; in computer vision some tasks now achieve over 99% accuracy.

Personal Protective Equipment (PPE) Detection

A computer-vision application that automatically identifies whether workers are wearing safety gear such as hard hats or vests.

Geospatial Imagery

Aerial or satellite images tied to geographic coordinates, used by insurers to assess property conditions during underwriting.

Underwriting

The insurance process of evaluating risk and determining appropriate coverage terms and premiums.

Claims Handling

The workflow an insurer follows to assess, validate, and settle an insurance claim, now often expedited by computer vision.

Recommendation Engine

An AI system that suggests actions or treatments (e.g., medical interventions) based on analyzed data and predicted outcomes.

Feedback Loop

The process of feeding a model’s outputs back into its training cycle to continuously improve accuracy and performance.

Data Cleansing

The rigorous process of identifying and correcting or removing inaccurate or inconsistent data to ensure high-quality inputs for AI models.

Computing Power

The processing capacity (CPUs, GPUs, etc.) required to train and run resource-intensive computer-vision algorithms.

Facial Recognition

A computer-vision technology that identifies or verifies individuals by analyzing facial features in images or video.

Surveillance

The monitoring of behavior or activities, often using technologies like facial recognition, raising privacy and ethical concerns.

Privacy

The right of individuals to control how their personal information and likeness are collected, used, and shared.

Model Drift

The degradation of an AI model’s performance over time as data patterns change, necessitating ongoing retraining.