Homepage » Amazon Web Services Solutions » Machine Learning » AI Solution Kit
AI Solution Kit
One stop to find various AI solutions for common use cases
Homepage » Amazon Web Services Solutions » Machine Learning » AI Solution Kit
AI Solution Kit
One stop to find various AI solutions for common use cases
Overview
AI Solution Kit provides a series of out-of-the-box AI features on the cloud, such as multilingual Advanced Optical Character Recognition (OCR), general natural language understanding (NLU), general object recognition, image super resolution, image similarity, and human image segmentation.
The solution provides RESTful API interfaces and can be seamlessly integrated with other services provided by Amazon Web Services.The solution provides two types of architecture: architecture based on Amazon Lambda and architecture based on Amazon SageMaker, which you can choose according to your needs.
-
Optical Character Recognition (OCR)
-
Face and Body
-
Image Understanding
-
Visual Content Production
-
Natural Language Understanding (NLU)
-
Optical Character Recognition (OCR)
-
Lite OCR
Recognize and extract Simplified Chinese, Traditional Chinese, numbers, alphabetical characters and symbols from images. Return the information such as text or coordinates.
Advanced OCR
Recognize and extract Simplified Chinese,Traditional Chinese, Vietnamese, Japanese, Korean, English, numbers, alphabetical characters and symbols. Return the information such as text or coordinates.
Custom OCR
Customize the OCR template, extract the structured text information in cards and tickets, and display the results in the key-value format.
Car License Plate
Recognize text on Chinese car license plate.
-
Face and Body
-
Face Detection
Detect the face in an image, and map the detected facial features and contour key point information to 64 vector coordinates.
Detect the face in an image, and map the detected facial features and contour key point information to 64 vector coordinates.Face Comparison
Compare two faces of the same person and return a confidence score of the similarity.
Human Attribute Recognition
Recognize the attributes of the human body in the image, and return the position coordinates and attribute analysis in each area, including the semantic information of 16 attributes.
Human Image Segmentation
Segment human bodies from background and return the alpha channel which is a color component representing the degree of transparency of a color.
-
Image Understanding
-
Image Similarity
Compare two images by calculating the cosine distance from the image feature vector and converting it into confidence, and return similarity score.
Object Recognition
Recognize objects in the image and return the region information and confidence score for each object. 300 types of objects are supported.
Pornography Detection
Detect pornographic image in three dimensions (normal, sexy, porn) and return confidence scores.
-
Visual Content Production
-
Image Super Resolution
Upscale the resolution and enhance details in the images, and support image enhancement.
-
Natural Language Understanding (NLU)
-
Text Similarity
Compare two Chinese words or sentences and return similarity score.
General NLU
Support a variety of Chinese text understanding tasks, such as text classification, sentiment analysis, extraction, and customizable labeling systems.
Benefits
You can find the code implementation of most common AI use cases. All solutions and applications included are vetted by Amazon Web Services.
You can find the code implementation of most common AI use cases. All solutions and applications included are vetted by Amazon Web Services.
The solutions automatically create RESTful APIs through Amazon API Gateway. After deploying the solution, you can send HTTP POST requests to use the AI functions. The deployment guides provides all API test methods (e.g. Postman) and offers sample codes.
The solutions automatically create RESTful APIs through Amazon API Gateway. After deploying the solution, you can send HTTP POST requests to use the AI functions. The deployment guides provides all API test methods (e.g. Postman) and offers sample codes.
All the solutions are open sourced. You can use the solutions and applications for free. If you have different use cases, you can take the source code as reference to make your own implementation.
All the solutions are open sourced. You can use the solutions and applications for free. If you have different use cases, you can take the source code as reference to make your own implementation.
Solution overview
The diagram below presents the architecture you can automatically deploy using the solution's implementation guide and accompanying Amazon CloudFormation template.
This solution architecture has two types of AI feature implementations: architecture based on Amazon Lambda and architecture based on Amazon SageMaker.
Architecture based on Amazon Lambda
1. The user or application sends an API request to the Amazon API Gateway. The request payload needs to contain information about the image or text to be processed.
2. Amazon API Gateway sends the incoming user request directly to the Amazon Lambda function.
3. The Amazon Lambda function returns the result to the invoker.
Architecture based on Amazon SageMaker
1. The user or application sends an API request to the Amazon API Gateway. The request payload needs to contain the image or text information to be processed.
2. Amazon API Gateway sends the request to Amazon Lambda (invoke endpoint) function.
3. Amazon Lambda invokes Amazon SageMaker Endpoint, which performs the inference process in Amazon SageMaker and returns the inference result (usually in JSON format data).
Architecture based on Amazon SageMaker
Architecture based on Amazon SageMaker
1. The user or application sends an API request to the Amazon API Gateway. The request payload needs to contain the image or text information to be processed.
2. Amazon API Gateway sends the request to Amazon Lambda (invoke endpoint) function.
3. Amazon Lambda invokes Amazon SageMaker Endpoint, which performs the inference process in Amazon SageMaker and returns the inference result (usually in JSON format data).
Architecture based on Amazon Lambda
Architecture based on Amazon Lambda
1. The user or application sends an API request to the Amazon API Gateway. The request payload needs to contain information about the image or text to be processed.
2. Amazon API Gateway sends the incoming user request directly to the Amazon Lambda function.
3. The Amazon Lambda function returns the result to the invoker.
Related content
Unlock the Power of Generative AI in the Cloud with Ease and Speed
Source code
