Inference using Mask-RCNN custom model ======================================== **Mask-RCNN** architecture can be used for multiple vision task. In this case it is used for **Instance Segmentation**. Direct result: * Binary raster (tree/non tree) Derived results: * Polygons vector layer * Bounding Boxes vector layer * Centroids vector layer =================== Input =================== For this example we will use this image of a silvopastoral system. .. csv-table:: Input image features :file: ../res/example_2.csv :align: center :header-rows: 1 .. image:: ../res/input_raster_png/example_2.png | The images can be downloaded `here `_. =================== Configuration =================== Input ------ * Select the appropiate **Input layer** from the dropdown menu. * For **Extent** select **Layer extent**. Output ------- * Select the appropiate **Output directory** and **Output name**. Processing ---------- * Go to the **Inference** tab * Select **Mask-RCNN** model from the dropdown menu * Select the desired **Result types**, the direct result type for this model is **Binary** * Press **Process** to start the inference process =================== Results =================== The resulting added layers depend on the selected **Result types** .. .. list-table:: .. :widths: 25 25 25 25 .. * - .. image:: ../res/results_png/example_1_grayscale.png .. - .. image:: ../res/2-3-2023-crop2.png .. - .. image:: ../res/2-3-2023-crop3.png .. - .. image:: ../res/2-3-2023-crop4.png Direct result: .. list-table:: :widths: 30 30 40 :header-rows: 1 * - Processing Result - Result Type - Result * - **RASTER LAYERS** - - * - **Direct result** - Binary - .. image:: ../res/results_png/example_2_binary.png * - Derived result - Grayscale - NOT AVAILABLE FOR THIS MODEL * - **VECTOR LAYERS** - - * - Derived result - Polygons - .. image:: ../res/results_png/example_2_polygons.png * - Derived result - Bounding Boxes - .. image:: ../res/results_png/example_2_bb.png * - Derived result - Centroids - .. image:: ../res/results_png/example_2_centroids.png