Improve accuracy of page detectors

[Witiko]

The low accuracies in the experimental notebook from #3 are worrying:

Page Detector	random	siamese	imagehash	vgg16	annotated
Accuracy	0.03%	3.95%	6.58%	61.84%	100.00%

The vgg16 page detector currently uses features produced by the last hiddent layer of a VGG16 model trained on ImageNet. Finetuning may not be an option, since our dataset is ill-suited for classification (too many document pages/classes, too few examples of each class). However, since our dataset is significantly different from ImageNet, we may have better luck using an earlier hidden layer of VGG16:

• Update the implementation.
• Optimize the parameters.
• Evaluate.

The siamese page detector uses position-dependent samples to normalize the input screen images, which may explain why the performance degrades on new document page and screen images (86% accuracy on training set versus 3.95% accuracy on test set):

• Update the implementation.
• Estimate performance using CV.
• Train and optimize the parameters.
• Evaluate.

If we manage to improve the siamese page detector, we may benefit from ensembling siamese with vgg16:

• Estimate performance using CV.
• Add ensemble page detector to video699.__main__.
• Evaluate.