I need help in coming up with a technique to determine when an object is not in the list of already classified objects. I believe openmax will solve this problem.
I am sorting parts that are the identical shape, but differ in the text written on them. Unfortunately the text is written in a circle, so I can’t simply read the text with OCR. (At least I couldn’t make it work.)
I have about 20 different labels on the parts I currently have. Some parts I have 1,000 pictures of, others I only have 100 pictures.
I discovered while I was sorting that there are actually more than 20 different labels, there are parts I didn’t know about that only occur once or twice in five hundred parts. I would like to be able to identify and handle parts like this separately.
I am using [login to view URL] to classify the images (parts). I get about a 96% accuracy of a part being identified correctly when I use two different classifiers to look at the same image and then check their agreement. I do this because I would rather reject parts than have them sorted incorrectly.
My images are 300x300.
I look at the weight of the prediction, and 95% of my predictions are almost certain, say .99999, so I haven’t been able to do the simple thing of just looking at the strength of the prediction.
In case this is relevant, the resnet50 classifier that I have trained might be overtrained. Within three epochs it is at like 99.7% accuracy with a val_accuracy of 98%.
99/99 [==============================] - 572s 6s/step - loss: 0.0138 - accuracy: 0.9967 - val_loss: 0.3574 - val_accuracy: 0.9796
Here is an article that shows what needs to be implemented.
[login to view URL]
The project milestone is when the software correctly identifies out-of-distribution images (similar images that the classifier has not been trained on).
20 freelancere byder i gennemsnit $210 timen for dette job
I'm a senior Python & ML developer and owner & founder of Dedeoglu Dev Company. Kindly send me a message to get in touch with me, Thanks, Yusuf.
Hi, I have +5 years of experience dealing with machine learning algorithms and worked on multiple projects in this field, I absolutely can do your project as you like. Please contact me to discuss more. Have a nice day