1. Colect images To train a model to predict objects, for this challenge I need images of boxes of different sizes, we also need a static background image since if we consider that a logistics company uses it, the detection of objects would have to be through of a camera that identifies each package to be sent. In this challenge we decided that the packages should be based on these approximate measurements, with which we created the test package images to train the model. size lenght width tall XXS 15 10 10 XS 20 25 12 S 30 20 12 M 30 24 20 L 42 30 23 Custom 42> 30> 23>
We collected 18 images per size, and 3 interactions were made between the different sizes. XXS
XS
S
M
L
2. Azure Custom Vision Object - Detection
2. Azure Custom Vision Object - Detection
2. Azure Custom Vision Object - Detection Tagging images with size for train the Model. Setup train with images
2. Azure Custom Vision Object - Detection Result of train we must Get a Good precisión , recall And mAP values for build a Prediction URL
2. Azure Custom Vision Object - Detection Prediction API to use in our Application
3. Create API to consume in Application ExpressJS The API Will be recibe a image and it will give us the size and Probabilty the images will save in the same Web App Service Azure and for privacity that Images could be save in Azure Blop Storage.
3. Create API to consume in Application ExpressJS Code for interfaces of the API
3. Create API to consume in Application ExpressJS Deploying API in Azure Web App
3. Create API to consume in Application ExpressJS Interfaces
3. Create API to consume in Application ExpressJS Interfaces
3. Create API to consume in Application ExpressJS Testing Result