Deep Active Learning Framework for Lymph Node Metastasis Prediction in Medical Support System

Table 1

Time consumption (in seconds) of different update methods under random selection strategy.

	Retrain	Mistake replay	Proposed	Queried only
PCam	291.42	145.23	29.23	29.76
MNIST	387.17	192.93	38.73	36.63
CIFAR10	380.19	188.20	37.07	37.12