Supporting Several Vision Tasks
Object detection can obtain the positions and sizes of multiple different objects in the image, and can be represented by bounding boxes.
Image classification classifies images into different categories or labels to distinguish different objects, scenes, or concepts.
How does SingTown AI Cloud Service work?
Uploading pictures of objects to be identified, requiring captureing from multiple angles in actual scenarios. The more pictures, the better.
Classify images or draw bounding boxes on images to identify target objects.
Use annotated pictures to train the neural network model in order to recognize target objects.
Download the trained neural network model to the device. It could run offline on the device and does not require an Internet connection.
Simplely to use. You only need to upload pictures, use the mouse to draw the bounding box. We supports shortcut keys for easy data annotation. And supports importing labeled data sets.
There is no need to install software, buy a GPU, learn AI knowledge, or program. It only takes you one click to start training.
SingTown AI Vision Module
SingTown AI Vision Module is specially designed for AI. It is a small size, high performance and NPU built-in vision module.
OpenMV Cam is an open source vision module. It can be programmed by Python and integrates many image algorithms.
More devices will be supported in the future.