From 14c7f926887941f70815759c8cbca121fb2fd5ca Mon Sep 17 00:00:00 2001 From: Jamie Mac Date: Mon, 15 Jul 2024 16:45:12 +0100 Subject: [PATCH] Update deep_learning_help.md Updated help values to have some info on data selectors and symbol managers. --- .../deep_learning_help.md | 26 ++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) diff --git a/src/rawDeepLearningClassifier/deep_learning_help.md b/src/rawDeepLearningClassifier/deep_learning_help.md index 3e14a2cd..6912fc34 100644 --- a/src/rawDeepLearningClassifier/deep_learning_help.md +++ b/src/rawDeepLearningClassifier/deep_learning_help.md @@ -106,15 +106,15 @@ If using a deep learning model from a supported framework then all transforms ar _An AnimalSpot, Ketos or other supported deep learning model will automatically create a list of transforms with the appropriate settings. These is no need to use the advanced pane but it is there in case users wish to change transform settings for some reason_ -### Default Models +### Exanmple Models -Default models are selectable from the menu button in the Deep Learning Pane. Default models are deep learning classifiers which are open source, known to be effective and have have been published in open access academic litrature; they are downloaded directly from a GitHub repository and then all associated settings are automtically applied. The default model selection pane also contains hyperlinks to the papers descirbing each model which will take users directly to the relvent website. +Example models are selectable from the menu button in the Deep Learning Pane. Default models are deep learning classifiers which are open source, known to be effective and have have been published in open access academic litrature; they are downloaded directly from a GitHub repository and then all associated settings are automtically applied. The default model selection pane also contains hyperlinks to the papers descirbing each model which will take users directly to the relvent website.

-_Default models can be downloaded. Default models are models which are published, open and have been known to work well amongst the bioacoustics community. More will be added to PAMGaurd over time. If you you would like your model to become a defualt model then drop PAMGuard support an email._ +_Example models can be downloaded. Example models are models which are published, open and have been known to work well amongst the bioacoustics community. More will be added to PAMGaurd over time. If you you would like your model to become a defualt model then drop PAMGuard support an email._ ## Running @@ -144,6 +144,26 @@ _An example click detection module output coloured by deep learning annotations. Other displays also show outputs from the deep learning module. Hovering over data units in the click display will, for example, show deep learning prediction values. The spectrogram will also show deep learning detections as translucent blue boxes (these must be selected in the right click menu). +### Symbol Manager + +The deep learning classifier has a symbol manager that is used throughtout different PAMGuard displays. The symbol manager can either display deep learning detections by the prediction values of a particular class or colour the output class with the highest prediciton class. Colouring by a prediction value requires the user to select which output class they wish to display - all detections are then coloured by the output prediction values for that particular class via a selectable colour map. If colour by class is selected then each detection is coloured by the class with the maximum prediction value. The colours for each class are selectable. + +

+ +

+ +_The options for the deep learning symbol manager_ + +### Data selector + +The deep learning classifier has a data selector which is utilsied thorughout PAMGuard (for example for exporting detections, generating clips, showing on displays). The data selector allows users to filter detections based on a minimum prediciton values for different output classes. Classifiers with multiple classes can have different minimum predicitons values assigned to each class. The user interface allows user to use a slider to dynimically change the minimum prediciton value. The lock button clamps the minimum prediction values of all enabled output classes together. + +

+ +

+ +_The options for the deep learning data selector_ + ### MATLAB The easiest way to export to MATLAB is to select the desired units in the time base display, right click and select the MATLAB icon. Data units will be exported to a .mat file as list of structures which is then saved to the clipboard. This file can be saved and then dragged into MATLAB to open.