Computer vision
Computer vision is the field of computing concerned with using computational methods to extract and interpret information from images, video, and other visual data.
Web Components in Action
Computer vision is the computational analysis of images or video to infer information about the physical world.
- How it is used
- Enables augmented-reality libraries to locate real-world surfaces where virtual objects can be placed.
- How deep it goes
- Explains feature points and detected surfaces conceptually, without detailing computer-vision algorithms or implementation.
- Examples
- Uses floors, tables, and walls as detected surfaces and an ARCore demonstration of virtual-object placement.
- What you can do after
- Evaluate/understand how ARKit and ARCore use camera-derived features to support virtual-object placement.
Read how Web Components in Action explains computer vision →
Inside AI
Computer vision is the field concerned with enabling computers to interpret and analyze visual images or video.
- How it is used
- Uses computer vision to contrast machine image recognition with human visual understanding and to assess a proposed medical application.
- How deep it goes
- Explains recognition through familiar comparisons and reported failures, but does not describe model architecture or training procedures in detail.
- Examples
- Uses cat recognition, COVID-19 chest-image diagnosis, and a one-pixel image alteration to make capabilities and limitations concrete.
- What you can do after
- Evaluate/understand image-recognition claims by separating successful pattern matching from robust interpretation and demonstrated clinical usefulness.
How Large Language Models Work
Computer vision is the field concerned with enabling computational systems to analyze and interpret visual data.
- How it is used
- Explains how transformer-based systems interpret images, detect objects, identify objects, and generate images.
- How deep it goes
- Traces image-to-vector patch extraction and reverse reconstruction, but stops short of detailed training procedures or mathematical derivations.
- Examples
- Uses image patches, RGB vectors, a 16 × 16 patch containing 768 values, and patch extraction and combination diagrams.
- What you can do after
- Evaluate/understand how images become transformer inputs and how transformer outputs become images.
Read how How Large Language Models Work explains computer vision →
Digital Twins in Action
Computer vision is the computational analysis of images or video to derive information about objects, activities, and environments.
- How it is used
- Turns visual records and video into machine-readable observations for a digital twin.
- How deep it goes
- Works through meter-photo preprocessing and webcam object detection, while leaving production event publishing and validation workflows unimplemented.
- Examples
- Uses an analog water-meter photograph, webcam-based room occupancy detection, and a piping and instrumentation diagram.
- What you can do after
- Apply it via Tesseract OCR to extract meter readings or via YOLOv8 to inspect objects in a webcam feed.
Build AI-Enhanced Web Apps
Computer vision is the field concerned with enabling computers to interpret and reason about visual data.
- How it is used
- Enables applications to interpret images for object recognition, visual question answering, captioning, and image-based responses.
- How deep it goes
- Details image submission, processing, comprehension, and response generation while deferring implementation details.
- Examples
- Uses broken-appliance diagnosis, image captioning, visual question answering, and cat-image classification.
- What you can do after
- Apply it via an LLM API that accepts image data and prompts for vision-based responses.
Read how Build AI-Enhanced Web Apps explains computer vision →
Investing for Programmers
Computer vision is the field of enabling computers to interpret and analyze visual information such as images and video.
- How it is used
- Uses visual data as an alternative information source for analyzing companies, deployments, and market-relevant activity.
- How deep it goes
- Describes object counting and model training, but stops short of implementation details, evaluation metrics, or algorithm design.
- Examples
- Counts car brands in street imagery and evaluates Tesla Vision as a potential LiDAR replacement.
- What you can do after
- Apply it via computer vision algorithms to extract vehicle counts and assess visual-data opportunities.
Read how Investing for Programmers explains computer vision →
Deep Learning with R, Third Edition
Computer vision is the field concerned with enabling computers to interpret and analyze visual data such as images and video.
- How it is used
- Frames visual understanding through deep-learning applications such as classification, segmentation, detection, OCR, and autonomous systems.
- How deep it goes
- Defines three foundational tasks and mentions specialized ones, while deferring detailed convolutional-network architectures and training methods.
- Examples
- Uses photo search, video-call backgrounds, autonomous driving, OCR, robotics, medical diagnosis, and dogs-versus-cats classification.
- What you can do after
- Apply convolutional neural networks and pretrained models to image-classification problems and related visual tasks.
Read how Deep Learning with R, Third Edition explains computer vision →
Build Your Own Robot: Using Python, CRICKIT, and Raspberry Pi
Software that enables computers to interpret and analyze visual information.
- How it is used
- Enables robots to inspect camera feeds, recognize targets, and guide movement toward desired objects or locations.
- How deep it goes
- Explains camera input, QR detection, decoding, video sharing, and robot search without detailing underlying vision algorithms.
- Examples
- Uses face detection, QR-tagged objects, Raspberry Pi camera feeds, and a searching robot as concrete examples.
- What you can do after
- Build a robot capability that detects and decodes QR codes, streams camera video, and searches for matching targets.
Read how Build Your Own Robot: Using Python, CRICKIT, and Raspberry Pi explains computer vision →
Related concepts
- image segmentationhas part · Image segmentation is a computer-vision task that partitions an image and assigns labels to its pixels or regions to delineate objects or meaningful areas.
- classifierhas part · A classifier is a predictive computational model that assigns inputs to one or more discrete classes or categories.
- object detectionhas part · Object detection is a computer-vision task that identifies object instances in visual data, assigns them categories, and estimates their locations.
- edge computinguses · Edge computing is a distributed computing approach that places computation and data processing near the users, devices, or sources and destinations of data.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.