Learn about how to use the "Apply Instance Segmentation" algorithm in the Trimble eCognition Software to apply an external Instance Segmentation model (TensorFlow SavedModel format with Mask Layer output). In order to use the "Apply Instance Segmentation" algorithm you first need to

  • load the model using the "load convolutional neural network" algorithm
  • followed by the "Apply Instance Segmentation" algorithm


In this example we use a model trained on the COCO 2017 dataset from the TensorFlow HUB (https://tfhub.dev/tensorflow/mask_rcnn/inception_resnet_v2_1024x1024/1). Here is a list of instances the model can detect --> https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/. Simply download it, unzip it and then follow the instructions in the video. You can use any image from your hard drive, or look for one online.

First an example of the results on some random images found on my hard drive. Second is an example of applying this model in eCognition Developer on a Trimble MX9 Mobile Mapping data set:

Instance_segmentation_example_COCO.gif     bordeaux_small.gif




Your Answer

Articles in this section

Need more help?

Enter Community Contact Support