Currently, I am using Keras for image recognition.
I've finished learning safely,
Among the training data, test data, and verification data,
I want to use two images at the same time in a verification program to determine what it is.
What if I use a shared layer? I received an opinion, but
I didn't understand even if I looked it up, so I'd like to ask you a question.
Is the shared layer something to use before learning?
Or is it something to use after learning is over?
I would appreciate it if you could tell me.
Answer # 1
The term shared layer is uncommon.
Probably it refers to weight sharing, but please check with yourself just in case.
Weight sharing is the sharing of some or all of the network parameters.
The main benefit of weight sharing is the reduction of parameters.
As far as I read the comments, I want to improve the classification accuracy with two images of the same class.
Put the images in order on the same network and take a majority vote on each classification result. The easiest.
Put the images in order on the same network, add each output, and then classify.
Both are done at the time of inference.
Start with something simple.
Only challenge complex networks if you are not satisfied with these results.
- python - image recognition using cnn keras multiple inputs
- python - how to input grayscale image to keras vgg16 model (imagedatagenerator)
- how to save the image ocr result file in python
- python - image is not saved even though there is no opencv error
- python - keras graph partition
- python - (tensorflow) please tell me how to save and output image data mnist
- python - (tensorflow/speech recognition) what kind of numerical conversion method is there to enable input of arbitrary wav file
- python image display error beginner, thank you
- python - about image deletion in django
- python - i want to find the height information of each pixel in the image
- python - tkinter photo image question
- python - i want to read an image and display it
- python - i want to design a cnn with 2 inputs and 1 output with keras
- python - i want to divide the read image into pixels and label them
- python - i want to create a contour image
- python - i want to display an image with pysimplegui, but an error occurs
- python - i want to load graph data created with pyplot as image data without saving it
- python - out of range of predict function used in image recognition (?)
- python - confusion matrix display in image recognition on cnn