Embedded vision combines cameras, edge or cloud computing, software, and artificial intelligence (AI) to enable systems to “see” and recognize objects. Intel offers a comprehensive portfolio of solutions to bring AI to life. This includes CPUs for general-purpose computing and computer vision, as well as Vision Processing Units (VPUs) for acceleration. Designed for use in a wide range of environments, computer vision systems can quickly recognize objects and people, analyze audio-demographic characteristics, inspect manufactured products, and much more. Computer vision uses deep learning to create neural networks that then guide the systems in image processing and analysis. Thoroughly trained computer vision models can be used for object recognition, people recognition, and even motion tracking.
Month: February 2023
Machine Learning (ML)
Machine learning (ML) is a collection of mathematical methods for pattern recognition. These methods identify patterns, for example, by breaking down datasets into hierarchical structures that minimize entropy. Alternatively, similarities between datasets are determined using vectors, and trained or untrained patterns are derived from them. Machine learning algorithms are indeed capable of solving many everyday or even highly specialized problems. In the practical work of a machine learning developer, however, problems frequently arise when either too little data is available or the data has too many dimensions. Entropy-driven learning algorithms such as decision trees become too complex when there are many dimensions and vectors.
Natural Language Processing (NLP)
The abbreviation NLP stands for Natural Language Processing and refers to techniques and methods for the computational processing of natural language. The goal is to enable direct communication between humans and computers. Natural Language Processing (NLP) seeks to capture natural language and process it computationally using rules and algorithms. NLP draws on various methods and findings from linguistics and combines them with modern computer science and artificial intelligence. The goal is to establish the highest possible level of communication between humans and computers through language. This is intended to enable both machines and applications to be controlled and operated through language.
Text-to-Speech
Speech synthesis involves artificially generating human speech and using it for communicative purposes. Speech synthesis enables human-machine communication. Technically speaking, speech synthesis involves the conversion and output of encoded data in the form of speech, whereby speech synthesis systems store and phonetically classify corresponding words or word fragments. Two methods are used in speech synthesis: the use of previously recorded natural speech in the form of phonemes, words, and sentence fragments, and the generation of synthetic speech from data and text, whereby the first method can be used with a limited vocabulary, while synthetic generation works with an unlimited vocabulary. Such speech synthesis systems are used in computer-telephony integration (CTI), screen readers, text-to-speech, …
Speech-to-text
Speech recognition refers to the ability of a machine or program to recognize spoken words and sentences and convert them into a machine-readable format.
Voice recognition is used, for example, in call forwarding, voice dialing, and voice search. Voice recognition should be distinguished from voice identification, which is a biometric method of identifying a person based on their voice.