I have the following evaluation metrics but I don't know how to analyze this one? Is my model performing poorly? What Average precision should I have to have a decent model? My model is trained to detect exposed body parts such as female and male genitalia, buttocks and female breast.
-
$\begingroup$ Are you treating your object detection problem as binary classification? In other words, you try to find if an object is or isn't something (or exist/doesn't exist), or you have many objects and you wan't to classify it in one of those categories? $\endgroup$– liakoyrasNov 16, 2022 at 16:13
-
$\begingroup$ @liakoyras my apologies for not adding further background to my model. My model is trained to detect exposed body parts such as female and male genitalia, buttocks and female breast so it is not a binary classification. $\endgroup$– Amben UchihaNov 17, 2022 at 10:42
-
$\begingroup$ So, you have one model that classifies into 5 categories (4 body parts and nothing)? $\endgroup$– liakoyrasNov 17, 2022 at 10:46
-
$\begingroup$ @liakoyras my train and test images does not have images that is safe or absent of any nudity so just 4 categories. $\endgroup$– Amben UchihaNov 17, 2022 at 10:49
-
$\begingroup$ I am asking these questions because I am trying to understand how your metrics are defined. Because average precision can either mean the mean of the precision for each class or the mean precision for all classes and all thresholds. $\endgroup$– liakoyrasNov 17, 2022 at 10:58