Algorithm 1.

Pseudo code of Proposed CNN-LSTM model

1:	Dataset ← X, Y = {y1, y2, y3, …, yn}
2:	Performs image pre-processing
3:	Image = cv2.resize (224 x 224), resize the image.
4:	Computing Newimage = image (extTop [1]: extBot [1]: extLeft [0]: extRight [0]), Cropping the image using extreme point calculation.
5:	Splitting the dataset into validation and training parts. Thirty percent for validation and 70% for training.
6:	CNN-LSTM ← Classifying and extracting the feature through deep learning models.
7:	F = (f1, f2,f3, …, fn) map the feature extraction vector into high dimensional space.
8:	for every epoch in the number of epochs do
9:	for every batch in the batch-size do
10:	x = model (F);
11:	Loss = cross_entropy (X, x), Calculate the loss
12:	Optimization and fitting function applied for validation and training of the model
13:	Compute the validation metrics: precision, accuracy, F1-measure, and recall
14:	end for
15:	end for
16:	return Results