Prepare the release for 5.15

[EE-LiuYunhao]

Summary

Adding the features come with PICO OS 5.15, including:

1. Readback APIs, which allowing your app to read back tensor content from SecureMR service
2. JavaScript operator, which allowing your app to run customized JavaScript codes together with your algorithm outputs

Additionally, we provide some more developer-friendly tools, including:

1. A inspector to check whether the machine-learning models are correctly converted and set up
2. A JSON-based format that serializes SecureMR pipelines as plain-text exchangeable file, and
3. A deserializer which can dynamically create and deploy SecureMR pipelines from a JSON file

You can refer to the README for the MNIST sample to regarding the details of the deserializer and the JSON format

Documentation enhancement

Add the documentation and sample codes for Pipeline::arithmetic method, regarding the support for

• T: transpose
• inv matrix inverse
• ^ exponential
• sin, asin, sinh, cos, acos, cosh, tan, atan, tanh: trigonometric functions

New samples

1. samples/mnist: demonstrating how to use the JSON-based serialization of the SecureMR pipelines
2. samples/readback: a basic sample to show how to use the readback API to readback data (e.g., camera data) as memory, or OpenGL/Vulkan texture
3. samples/rubic_cube: a rubic cube solver using the readback API
4. samples/whackamole: a Whack-A-Mole game demonstrating the usage of JavaScript along side body-tracking algorithm to add fun to the MR app
5. samples/model_inspect: demonstrating how to use the model inspection tool to check the model conversion and setting-up.