Do you ever wish that you could train an AI model to identify objects using computer vision? Well, IBM developer Nicholas Bourdakos’ recently created Code Pattern would allow you to do just that!
The Code Pattern is being called “Create a real-time object detection app using Watson Machine Learning”. It depends on Cloud Annotations and serves as a template on which an AI model can be trained so it can detect any and every object. Data labeling has also been made quite easy, thanks to a new tool.
Bourdakos cites someone doing emoji classification as an inspiration behind his project.
Bourdakos and IBM also shared a tutorial that can help developers in understanding the Code Pattern and Cloud Annotations and how to use them.
Developers are required to build an iOS app that would allow them to use their custom-trained models in order to recognize objects. For storing the labeled data, an IBM Cloud Object Storage instance is required. Once the data is ready, developers can learn how to start Watson Machine Learning instance to train their custom models on world-class GPUs. The Core Machine Learning model can be plugged into the iOS application, once the training process has been completed.
Majority of this entire process is quite easy as in reality, an AI model doesn’t need to be trained from scratch, but rather it requires re-training. IBM has already trained the algorithms on countless images to make it learn various computer vision skills. So, all developers have to do is make the AI detect the images they are “annotating”.
The best part about this Code Pattern is that it is useful for both Highly Skilled Developers as well as aspiring ones with minimal Machine Learning skills.
So, in case you are a beginner looking to learn Machine Learning in the most fun way imaginable, or want to have your own object recognition AI models, Bourdakos’ project and Cloud Annotation should be your go-tos. Additionally, any other IBM Code Pattern can prove to be helpful as well.
Read next: A new AI is Using News Outlets to Produce Fake Article and it is quite Concerning!
The Code Pattern is being called “Create a real-time object detection app using Watson Machine Learning”. It depends on Cloud Annotations and serves as a template on which an AI model can be trained so it can detect any and every object. Data labeling has also been made quite easy, thanks to a new tool.
Bourdakos cites someone doing emoji classification as an inspiration behind his project.
Bourdakos and IBM also shared a tutorial that can help developers in understanding the Code Pattern and Cloud Annotations and how to use them.
Developers are required to build an iOS app that would allow them to use their custom-trained models in order to recognize objects. For storing the labeled data, an IBM Cloud Object Storage instance is required. Once the data is ready, developers can learn how to start Watson Machine Learning instance to train their custom models on world-class GPUs. The Core Machine Learning model can be plugged into the iOS application, once the training process has been completed.
Majority of this entire process is quite easy as in reality, an AI model doesn’t need to be trained from scratch, but rather it requires re-training. IBM has already trained the algorithms on countless images to make it learn various computer vision skills. So, all developers have to do is make the AI detect the images they are “annotating”.
The best part about this Code Pattern is that it is useful for both Highly Skilled Developers as well as aspiring ones with minimal Machine Learning skills.
So, in case you are a beginner looking to learn Machine Learning in the most fun way imaginable, or want to have your own object recognition AI models, Bourdakos’ project and Cloud Annotation should be your go-tos. Additionally, any other IBM Code Pattern can prove to be helpful as well.
Read next: A new AI is Using News Outlets to Produce Fake Article and it is quite Concerning!
