How to visualize the prediction? about fcn.tensorflow

[xiaoxionglin]

You have to train it first. The code does not contain the trained model.

[xiaoxionglin]

That's the model for object recognition. This program trains on the basis of that.

[Nii-co]

Hi Dean, got the same issue here, and I see the code downloaded a model from http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat, I thought it was the trained model for this project, but looks like it's not? So we need to trained from the beginning first?

[Nii-co]

Thanks for explanation, but why do we need an object recognition model? As I know FCN itself can do the recognition work independently, or maybe this extra model is for a better performance?
By the way, if my requirement is just to find a bottle in the image, and suppose this object recognition model can't recognize bottles, do I need to retrain the object recognition model first?

[ChinmayKalyankar]

@lxx74n74n can you help me how to train a model using this code...?
I mean what all things I have to download before doing it and which commands to fire ??