Scenarios

Object detection

Object detection identifies the coordinates of multiple objects in an image and provides their classification labels. Common scenarios for object detection include the following:

• Video detection: Pinpoint the location of objects in each frame of a video. For example, an in-vehicle system can intelligently detect its surroundings, including passing vehicles, pedestrians, lane lines, and traffic lights. This provides the driver with safety assistance, such as following distance warnings, lane departure warnings, traffic light monitoring, and advance lane change reminders.

• Image recognition: Improve recognition accuracy by trimming the detected objects from an image before performing recognition.

• Object localization: Classify and add labels to large volumes of images.

Image classification

Image classification categorizes images based on their content. EAIS supports these inference scenarios with flexible configurations and extensive network access.

Natural language processing

It supports operations on characters, words, sentences, and passages, such as input, output, detection, analysis, understanding, and generation. It provides the following:

• Content recommendation: Extract key semantic information using techniques such as keyword extraction and short text similarity. This enables you to accurately match semantically similar content to quickly build recommendation applications.

• Translation: Analyze and translate text to help users communicate across different languages.

• Public opinion analysis: Identify breaking events and public opinion trends by discovering topics and analyzing sentiment, hot spots, and trends from multiple dimensions. This provides a timely and complete view of public opinion.

• AI chat systems: Use techniques such as Chinese tokenization, short text similarity, and named entity recognition (NER). These techniques help address challenges in Q&A pairing, dialogue management, corpus mining, and knowledge base construction.