When an admin / expert detects an object, we assume that it is good. When a general user selects an object, we cannot confirm that it is good; therefore, we should create an interface whereby experts can go through objects detected by users and confirm that they are good.
We should only use good images in our testing data. Pursue after issues #3, #4, #5.
When an admin / expert detects an object, we assume that it is good. When a general user selects an object, we cannot confirm that it is good; therefore, we should create an interface whereby experts can go through objects detected by users and confirm that they are good.
We should only use good images in our testing data. Pursue after issues #3, #4, #5.