ENG [#VS2015] Hello Face: #FaceAPIs in a Console App (2)

Hello!

New post for Azure Machine Learning Face APIs series

  1. Face APIs in Azure
  2. Hello Face: Face APIs in a Console App

After the setup of our Azure environment, we now can use Face APIs. Next step is to download the SDK and take a look at the examples. At this time the SDK contains examples for both, .Net and Android. Within the .Net sample there is a WPF app which consumes a PCL which is responsible for making the calls to the Machine Learning Face API services.

Clipboard03

The good thing about this model is that the PCL is easily portable to other projects. The following steps show how consume Face API services in a console app, using the PCL included in the SDK.

1. Add a new console application in the solution

2. Add a reference to ClientLibrary

3. Then in the Main we need to define a variable with our subscription key, and for this example we will open a local image to be processed

4. The following function displays the steps necessary for processing an image

Clipboard05

5. And the result is similar to the following one

Clipboard07

6 That’s it !

As always, can always download the complete sample from GitHub https://github.com/elbruno/ProjectOxford

Greetings @ Home

/El Bruno

Face APIs SDK, http://www.projectoxford.ai/sdk

[#AZURE] Hello Face: #FaceAPIs en una Console App (2)

Hola!

Nuevo post para la serie de Face Api en Azure

  1. Face APIs en Azure
  2. Hello Face: Face APIs en una Console App

Una vez configurado nuestro entorno de Azure para poder consumir Face APIs, lo siguiente es descargar el SDK y darle un vistazo a los ejemplos del mismo. En este momento el SDK contiene ejemplos para .Net y Android. Dentro del ejemplo para .Net hay una app WPF de ejemplo que consume un PCL que es el encargado de hacer las llamadas a los servicios de Machine Learning de Face API.

Clipboard03

Lo bueno de este modelo, es que el PCL es fácilmente portable a otro tipo de proyectos. Los siguientes pasos muestran como consumir los servicios de Face API en una app de Consola, utilizando el PCL del SDK.

1. Agregar una nueva aplicación de consola a la solucíón

2. Agregar una referencia a ClientLibrary

3. Luego en el Main debemos definir una variable con nuestra subscription key, y para este ejemplo abriremos una imagen local para procesarla

4. La siguiente función muestra los pasos necesarios para procesar una imagen

Clipboard05

5. El resultado es similar al siguiente

Clipboard07

6. Listo !!!

Y el ejemplo, como siempre se puede bajar desde GitHub https://github.com/elbruno/ProjectOxford

Saludos @ Home

/El Bruno

Face APIs SDK, http://www.projectoxford.ai/sdk

ENG [#AZURE] Face APIs with #Azure (1) #MachineLearning

Hello!

Every time I perform a Coding4Fun session, I always take the opportunity to talk a little about the progress in the process of face detection, facial recognition and detection of emotions, etc. If you like Azure, now is a great time to start testing something for this topic, since using Machine Learning experiments, there are a number of features available to perform these actions.

In this series of posts I’ll show you how to configure Azure for having an instance of Face APIs, how to active and publish it as a service and finally as consume it from an .Net app.

You should start adding a Face API instance from the Azure Machine Learning Market Place Gallery . Access the Market Place and add search for Face APIs .

Clipboard01

The wizard is fairly simple, and it’s free. For now is only available in the West US region, although that doesn’t affect us much.

Clipboard02

Once created the instance, it will appear in our list of items for the Market Place section. The next step is very important, since it is where to generate the key that we will identify to use this service from our apps. We must Access the portal Face APIs from the option “Manage”

Clipboard02

You will find your primary and secondary keys and also the option to regenerate them

Clipboard02

And that’s it! We already have our Azure environment ready to use Face APis. I’ll write later about the Face APIs capabilities, however the information and the SDK can be found at the official website of the Project Oxford , where besides Face APIs, there are APIs for Speech Recognition and Computer Vision. Come on, that is a place to have some serious fun  ;)

Greetings @ Home

/El Bruno

Project Oxford, http://www.projectoxford.ai/face

Face Api Machine Learning, http://gallery.azureml.net/MachineLearningAPI/b0b2598aa46c4f44a08af8891e415cc7

[#AZURE] Face APIs con #Azure (1)

Hola!

Desde hace un tiempo en las sesiones de Coding4Fun siempre aprovecho para hablar un poco de lo que han avanzado los procesos de detección de rostros, reconocimiento facial, detección de emociones, etc. Si te gusta Azure, ahora es un gran momento para comenzar a probar el mismo, ya que utilizando experimentos de Machine Learning, hay disponible una serie de features para realizar este tipo de acciones.

En esta serie de posts mostraré como configurar Azure para tener activa una instancia de Face APIs, como publicar la misma como servicio y finalmente como consumir la misma desde una app .Net.

Lo primero es consumir Face API, desde la gallery de Azure Machine Learning. Para esto accedemos al MarketPlace y agregamos Face APIs.

Clipboard01

El asistente es bastante simple, y además es gratis. Por ahora solo está disponible en la region de West US, aunque eso no nos afecta mucho.

Clipboard02

Una vez creado el mismo, ya lo podemos ver en la lista de elementos que tenemos en nuestra cuenta de Azure. El siguiente paso es bastante importante, ya que es donde generaremos la key que nos identificará para utilizar este servicio desde nuestras apps. Para esto debemos acceder al portal de Face APIs desde la opción “Manage”

Clipboard02

En el mismo nos encontraremos con nuestras claves primarias y secundarias, además de la opción de regenerarlas

Clipboard02

Y listo !!! Ya tenemos nuestro entorno de Azure preparado para utilizar Face APis. Sobre las capacidades que posee estas APIs escribiré más adelante, sin embargo toda la información y el SDK se puede encontrar en la página oficial del Project Oxford, donde además de Face APIs, hay APIs para Speech Recognition y Computer Vision. Vamos, que es un sitio para pasárselo en grande ;)

Saludos @ Home

/El Bruno

Project Oxford, http://www.projectoxford.ai/face

Face Api Machine Learning, http://gallery.azureml.net/MachineLearningAPI/b0b2598aa46c4f44a08af8891e415cc7

[#AZURE] Azure #MachineLearning free eBook

Hello!

Today is a short post, now is time for some running series and my legs are not in the mood. A few days ago I wrote a post with some options to start with Machine Learning. One of the options was Azure Machine Learning.

Today I read that the our MSPress friends have released a free book on this topic,

3312.9780735698178-FB_thumb_043387F7

The book covers the most important things for those who we are beginning to learn this amazing world. This is the chapter list:

  • Chapter 1, “Introduction to the science of data”
  • Chapter 2, “Getting started with Azure Machine Learning”
  • Chapter 3, “Using Azure ML Studio”
  • Chapter 4, “Creating Azure ML client and server applications”
  • Chapter 5, “Regression analytics”
  • Chapter 6, “Cluster analytics”
  • Chapter 7, “The Azure ML Matchbox recommender”
  • Chapter 8, “Retraining Azure ML models”

I jumped directly to the algorithms section (Chapter 2) and is quite well explained. It contains everyday samples and they are similar to what we can be found in our daily basis.

Then I jumped to Chapter 3, which is something I like a lot: a step by step tutorial on how to make a Hello world with AzureML. I liked: is simple and direct. A good tutorial that also relies on all the online material that is already available to learn more about AzureML.

Well, here I left it and I have study material for the next few days.

Saludos @ Home

/El Bruno

Source: http://blogs.msdn.com/b/microsoft_press/archive/2015/04/15/free-ebook-microsoft-azure-essentials-azure-machine-learning.aspx

[#AZURE] Libro electronico gratis de Azure #MachineLearning (o #AzureML como le decimos en casa)

Hola !

Post corto de viernes que ahora me tocan series y las piernas no están de humor para lo que sigue. Hace unos días escribí un post con algunas opciones para comenzar con Machine Learning. Una de las opciones era Azure Machine Learning.

Hoy veo que los amigos de MSPress han liberado un libro gratis sobre el tema, al que me tocará echarle un vistazo lo antes posible,

3312.9780735698178-FB_thumb_043387F7

El libro cubre lo más importante, para aquellos que estamos comenzando a conocer este mundo. Esta es la organización del libro

  • Chapter 1, “Introduction to the science of data”
  • Chapter 2, “Getting started with Azure Machine Learning”
  • Chapter 3, “Using Azure ML Studio”
  • Chapter 4, “Creating Azure ML client and server applications”
  • Chapter 5, “Regression analytics”
  • Chapter 6, “Cluster analytics”
  • Chapter 7, “The Azure ML Matchbox recommender”
  • Chapter 8, “Retraining Azure ML models”

He saltado directamente a la parte de algoritmos (Chapter 2) y la verdad es que lo explican bastante bien. Los ejemplos son cotidianos y se asemejan a lo que podemos encontrarnos en nuestro día a día.

Luego llegamos al Chapter 3, que es el que más me suele interesar: un paso a paso para hacer el Hola Mundo con AzureML. Me ha gustado, es sencillo y directo. Un buen tutorial que además se apoya en todo el material online que ya hay disponible para conocer más sobre AzureML.

Pues eso, aquí lo he dejado y ya tengo material de estudio para los próximos días.

Saludos @ Home

/El Bruno

Source: http://blogs.msdn.com/b/microsoft_press/archive/2015/04/15/free-ebook-microsoft-azure-essentials-azure-machine-learning.aspx

[#IOT] Glitter problems at home, Amazon #Dash and Azure #MachineLearning

Hello!

A couple of days I was painting with my 5-year-old girl and we realized that we had been out of glitter. As a parent this is one of the moments more happy complicated ever. Glitter is something that girls love to use in their art work.

Clipboard02

However, I discovered in the bad way that GLITTER is impossible to clean. If you slightly dropped a little and you try to use some cloth for cleaning, glitter is multiplied and increasingly everywhere. As always, someone see this as a problem, others finds here an opportunity. I find an online company, and they can send “glitter bombs” to your enemies, to basically ruin their day (or week). https://shipyourenemiesglitter.com/

Clipboard05

Going back to my problem, at that time, it would been very useful to have active an Amazon Dash Button associated with glitter as a product. The idea of this button is very simple: you put a button associated with a product in your home. Once that product is about to end, you press the button and this product it is added automatically to your Amazon shopping list. Later you can close the transaction from your Smartphone or laptop, and you’ll have it the next day at home.

When this was presented a couple of weeks ago, I listened to all different types of opinions. As always, the “haters” with opinions like

Does that mean I’ll have my house full of “buttons”?; When I have to buy several things I have to do clicks all over my house? Get out … for me it is much more easy to make online purchase from my computer.

And the last statement was the answer to their problem: for this user profile is easier to make the purchase directly online. So far, no one forces them to use Dash Buttons.

In my case, after years of training I finally get to a point where my children can prepare their breakfast themselves. That means that they are looking for cereals, milk and cocoa; and while I try to not help them, they are responsible for their breakfast. But … this doesn’t avoid me to have a conversations like this one with Valentino (7 years old)

  1. Valentino: Bruno, no more cereals
  2. Bruno: No problem, today I’ll get more. Do you have enough for today?
  3. Valentino: No, we get our of cereals, yesterday morning
  4. Bruno: Yesterday? Next time, let me know before
  5. Valentino: Before what?
  6. Bruno: There is no more cereals
  7. Valentino: Ahh Ok. Bruno, no more grain
  8. Go To 1

The conversation can be quite surreal with a 7-year-old child, and I must have a very high level of patience to not spend my next 5 years in prison. Perhaps a Dash Button, can help me to make this type of situation more bearable. If my children can handle their own stock of “breakfast stuff”, then is an excellent time to distribute responsibilities.

In the case of Amazon Dash, they have also led him to a higher level with 2 interesting options.

In addition to the button, there is a smart scanner that allows you to do something similar: scan the barcode of the product you want to purchase. This allows a very agile and fast build of the shopping list, since you are looking for stuff already in your home, you don’t have to leave home.

Also, the Dash platform has a dedicated space for Makers. This means that you can add this “button” in your own devices or even extend the platform. This is where the ability to innovate in hardware becomes important, or you can also take advantage of the power of Azure Machine Learning. For example, you could work on this basis:

  • I have an experiment in Machine Learning that analyzes my purchases
  • There is another experiment dedicated to analyze weather for the next 7 days
  • These experiments have detected every 5 weeks (approximately) I have to refill my glitter “stock”
  • In addition, crossing this data, the experiments can predict that when there is bad weather, I used more glitter
  • This is normal, when there is bad weather we can get out, so is time for home activities
  • Another experiment can analyze the weather and because it will rain in the coming days and my stock of “glitter” is low, it can suggest or add this product into my shopping list

I personally think that these examples of IoT (Internet of Things), are what really make a smart home or a setting’s value.

Connect a temperature sensor to a Raspberry Pi or an Arduino is fine, connect all the points for something of value, but the aim to achieve.

Clipboard03

Important: I can’t get away from my Scrum / Agile past, I always have to think to add value  ;)

/El Bruno

References

Glitter Bombs, https://shipyourenemiesglitter.com/

Amazon Dash Button, https://www.amazon.com/oc/dash-button?ref_=B00Q3E4UBW

Amazon Dash for Makers, https://www.amazon.com/oc/dash-replenishment-service

Azure Machine Learning, http://azure.microsoft.com/en-us/services/machine-learning/

Hardware is the future, with VisualStudio2013, http://elbruno.com/2014/05/24/opinion-hardware-is-the-future-with-visualstudio2013-of-course/

Internet of Things, http://en.wikipedia.org/wiki/Internet_of_Things