High Precision Dimensional Measurement with Convolutional Neural Network and Bi-Directional Long Short-Term Memory (LSTM)

. 2019 Dec 2;19(23):5302. doi: 10.3390/s19235302

Algorithm 1. The Process of the Transformation Operation

1. All the outputs from VGG-16 are denoted as

V_{i}

, the form is

(w, h, c)

, where

w

h

is the width and height of the feature map, and

c

is the number of channel. Expand the dimension of

V_{i}

at first axis, the new form of

V_{i}

(1, w, h, c)

2. Concat all the

V_{i}

at first axis, the output is denoted as

V

, and the form is

(N, w, h, c)

3. Expand the dimension of

V

, the new form is

(1, N, w, h, c)

4. Compress the width, height, and channel to one dimension, the form of the final output is

(1, N, w * h * c)