Morpho Deep Segmentation™
“Morpho Deep Segmentation™” is a generic API library for semantic segmentation. This is based on the inference engine “SoftNeuro®” and provides an easy-to-use API for high-speed operation, as well as common image-processing functions required for segmentation.
“Morpho Deep Segmentation” is a library of APIs that enables you to perform semantic segmentation tasks easily and quickly. Various pre-processing and post-processing routines common to segmentation tasks are integrated, allowing developes to focus on model development and application development. It also offers common visualizations as an API in order to accelerate model development and application debugging.
For example, in addition to outputting binary images for each category, you can also superimpose the confidence level of each category on the input image or output a color-coded image showing the categories.
Msodel developed using common frameworks such as Keras and Tensorflow, etc. are converted to “SoftNeuro” format for inference.
In addition to models provided by Morpho, models developed by customers can also be used.
It can also run on a variety of platforms, including x86_64 Windows and Linux, as well as ARM-based Linux and Android.
Note: In addition to Morpho Deep Segmentation, some other products such as Morpho Portrait Bokeh and Morpho Semantic Filtering, too, perform image segmentation using conventional image-processing technciques.