Skip to main content

Table 2 Features of the implemented neural network

From: Ultrasound characteristics of follicular and parafollicular thyroid neoplasms: diagnostic performance of artificial neural network

Layer

Shape

Number of Parameters

Activation Function

Input layer

(7,1)

0

-

Dense layer

[(7,50), (50)]

450

ReLu

Dense layer

[(7,250), (250)]

12,750

ReLu

Dense layer

[(7,100), (100)]

25,100

ReLu

Dense layer

[(7,2), (2)]

101

Sigmoid

Output

(2,1)

0

-

  1. Shape: the numbers represent the data vectors, ReLU: Rectified linear unit
Back to article page

Thyroid Research

ISSN: 1756-6614