A Novel Hybrid Approach for Classifying Osteosarcoma Using Deep Feature Extraction and Multilayer Perceptron

. 2023 Jun 18;13(12):2106. doi: 10.3390/diagnostics13122106

Algorithm 1 Proposed Algorithm

1:
$E_{p} \leftarrow Number of Epochs$
2:
$W \leftarrow Transfer Learning Model Parameter$
3:
$η \leftarrow Learning Rate$
4:
$b_{s} \leftarrow Batch Size$
5:
$D \leftarrow Osteosarcoma Dataset$

$Output : The assessment metrics on the test dataset .$

$Dataset Prepossessing :$
6:
$X_{t r a i n} \leftarrow p r e p o s s e s s i n g (D)$
7:
$X_{t e s t} \leftarrow p r e p o s s e s s i n g (D)$
8:
$Initialise TL Models (VGG 16, VGG 19, ResNet 50$

$Xception, DenseNet 121)$

$Feature Extraction :$
9:
for $local epoch e_{p} \leftarrow from 1 to E_{p}$ do
10:
for $b_{s} = (x_{s}, y_{s}) \in random batch from X_{t r a i n}$ do
11:
$Optimise model parameters$
12:
$W_{s} \leftarrow W_{s} - η (Δ (L (W_{s}; b_{s})))$
13:
$f_{t r a i n} \leftarrow C o m p u t e F e a t u r e s (W_{s}, X_{t r a i n}, 1024)$
14:
end for
15:
end for

$Feature Selection :$
16:
$f_{b e s t} \leftarrow D T - R F E (f_{t r a i n}, 900)$

$Osteosarcoma Tumor Classification :$
17:
$T r a i n e d M o d e l \leftarrow M L P (f_{b e s t}, y_{t r a i n})$
18:
$P r e d \leftarrow T r a i n e d M o d e l (X_{t e s t})$
19:
$E v a l u a t i o n m e t r i c s \leftarrow C o m p u t e M e t r i c s (P r e d, y_{t e s t})$