Hi, I'm an eighth grader, who is taking part in a coding competition. I am building an app for the students of my school to recognize 4 species of flora found in our school. This is to help city-bred children become more involved in nature. On one screen of my app, I am creating a flora recognizer. I trained my model on Teachable Machine, and using the teachable machine extension, I connected it to my app. However, there are two issues: 1) It doesn't recognize the species correctly consistently and 2) It doesn't show me one result; it ends up showing the wrong percentage for all the species making it difficult for new users to understand which species it is. The deadline for the competition is 30 November (Indian time) so it would really help if someone could guide me on this.
This might be an issue of your training set or the camera image being unclear. Perhaps the images of the different plants are too similar to each other or the camera plant image has a cluttered background A tool like this cannot be 100% correct. Did you do the tutorial that accompanied the extension link?
Wrong percentage? The percentages are the algorithm's best guess based on the information you provide.
is an image from the tutorial. Are your results similar.? I expect the percentages do not necessarily add up to 100% in your case.
You can try improving and increasing your training set of images.
