Tetra-ConvBiNet: MRI-based brain tumor classification using distributed tetra head attention-enabled convolutional bidirectional network

Chavan, Aasha Mahesh; Mane, Vanita

doi:10.2478/ijssis-2026-0005

Figures & Tables

Block diagram of the SEnO-DTCBiNet framework. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

SE attention. SE, squeeze and excitation.

Structural diagram of CC. CC, criss-cross.

Architecture of DTCBiNet. DTCBiNet, distributed tetra head attention-based convolutional bidirectional network.

Dataset visualization. (A) BraTS 2020 dataset, (B) BT dataset, and (C) BraTS 2018 dataset. BT, brain tumor.

Image results based on the BraTS 2020 dataset. ROI, region of interest.

BT dataset-based image results. BT, brain tumor; ROI, region of interest.

BraTS 2018 dataset-based image results. ROI, region of interest.

SEnO-DTCBiNet model performance on the BraTS 2020 dataset. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

Performance of the SEnO-DTCBiNet model using the BT dataset. BT, brain tumor. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

Performance of SEnO-DTCBiNet model using the BraTS 2018 dataset. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

SEnO-DTCBiNet model performance comparison on the BraTS 2020 dataset. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

K-fold analysis on the BraTS 2020 dataset.

Comparison of the SEnO-DTCBiNet model based on the training percentage using the BT dataset. BT, brain tumor. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

K-fold analysis on the BT dataset. BT, brain tumor.

Comparison of the SEnO-DTCBiNet model based on training percentage using the BraTS 2018 dataset.

Comparison of the SEnO-DTCBiNet model on the real-time dataset concerning training percentage. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

K-fold analysis on the real-time dataset.

ROC analysis. (A) BraTS 2020 dataset, (B) BT dataset, and (C) BraTS 2018 dataset. BT, brain tumor.

Time complexity analysis. (A) BraTS 2020 dataset, (B) BT dataset, and (C) BraTS 2018 dataset. BT, brain tumor.

Convergence analysis. SSA, Sparrow Search Algorithm.

Computational complexity analysis in terms of FLOPS. FLOPS, floating-point operations per second.

Ablation study for SEnO-DTCBiNet model components. SEnO-DTCBiNet, sonar energy optimized distributed tetra head attention-based convolutional bidirectional network.

Ablation study for attention mechanisms. CC, criss-cross; DTHA, distributed tetra head attention; SE, squeeze and excitation.

T-test analysis based on the BraTS 2020 dataset

Methods	T-test analysis

	Accuracy		Precision		Recall		F1-score

	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic
EDN-SVM	0.12	2.20	0.10	2.39	0.16	1.87	0.13	2.11
Dense-CNN	0.09	2.42	0.09	2.53	0.12	2.19	0.10	2.36
CJHBA-DRN	0.11	2.21	0.11	2.27	0.12	2.13	0.12	2.19
PatchResNet	0.11	2.24	0.12	2.14	0.10	2.40	0.11	2.28
BA-MCBM	0.13	2.11	0.11	2.25	0.18	1.76	0.13	2.04
2D-CNN-CAE	0.13	2.07	0.11	2.22	0.19	1.70	0.14	2.00
DE-MCBM	0.14	2.02	0.13	2.11	0.16	1.86	0.14	1.98
SSA-DTCBiNet	0.12	2.20	0.11	2.22	0.12	2.16	0.12	2.19
SPO-MCBM	0.11	2.21	0.12	2.17	0.11	2.28	0.11	2.22
SEnO-DTCBiNet	0.08	2.58	0.11	2.24	0.06	3.00	0.07	2.71

Results of the SEnO-DTCBiNet and baseline models based on training percentage

Methods/metrics		DE-MCBM	Patch ResNet	EDN-SVM	Dense-CNN	2D-CNN-CAE	CJHBA-DRN	BA-MCBM	SSA-DTCBiNet	SPO-MCBM	SEnO-DTCBiNet
BraTS 2020 dataset	Precision (%)	91.11	87.90	85.33	83.42	89.36	87.33	90.22	90.67	95.49	97.30
	Recall (%)	96.24	93.73	92.92	91.90	95.48	93.37	96.07	96.15	97.77	97.88
	Accuracy (%)	93.67	89.89	87.40	86.19	90.72	88.81	91.24	92.46	96.53	97.17
	F1-Score (%)	93.61	90.72	88.96	87.45	92.32	90.25	93.05	93.33	96.62	97.25

BT dataset	Precision (%)	94.39	89.04	86.73	85.09	90.12	86.74	90.57	95.55	96.71	96.86
	Recall (%)	96.15	93.14	92.51	91.74	94.74	92.97	95.83	97.33	98.52	98.72
	Accuracy (%)	94.30	91.35	89.05	88.57	91.78	91.13	91.82	96.11	97.91	98.48
	F1-Score (%)	95.263	91.05	89.53	88.29	92.37	89.75	93.128	96.43	97.61	97.98

BraTS 2018 dataset	Precision (%)	93.14	91.65	89.95	89.33	92.76	90.09	93.08	93.49	93.89	94.16
	Recall (%)	94.75	92.82	90.60	90.15	93.01	92.72	94.06	95.45	97.03	97.42
	Accuracy (%)	95.59	94.05	89.91	88.16	94.13	91.21	94.30	95.69	96.80	97.76
	F1-Score (%)	93.94	92.23	90.27	89.74	92.88	91.39	93.56	94.46	95.43	95.76

Real-time dataset	Precision (%)	94.80	93.27	91.62	90.45	93.96	91.73	94.24	95.29	95.31	95.99
	Recall (%)	95.88	94.57	92.93	92.56	94.88	93.85	95.13	96.89	97.03	97.10
	Accuracy (%)	95.16	93.70	92.06	91.15	94.27	92.44	94.54	95.82	95.88	96.36
	F1-Score (%)	95.34	93.92	92.27	91.49	94.42	92.78	94.68	96.08	96.16	96.54

Real-time dataset-based statistical analysis

Methods/metrics		Dense-CNN	EDN-SVM	CJHBA-DRN	Patch ResNet	2D-CNN-CAE	BA-MCBM	DE-MCBM	SSA-DTCBiNet	SPO-MCBM	SEnO-DTCBiNet
Accuracy	Best	88.00	90.46	91.23	92.92	93.66	94.52	95.38	95.52	95.58	96.48
	Mean	84.17	85.45	86.68	87.59	88.59	89.34	90.31	90.90	91.68	93.11
	Variance	6.30	11.08	11.89	14.74	16.74	15.91	13.40	12.30	9.84	6.77
	Standard deviation	2.51	3.33	3.45	3.84	4.09	3.99	3.66	3.51	3.14	2.60

Precision	Best	89.65	90.03	90.83	92.41	93.50	93.54	94.39	94.55	94.61	95.78
	Mean	84.94	85.55	86.29	87.12	88.20	88.70	89.83	90.54	90.71	92.31
	Variance	9.44	9.69	10.04	13.57	17.21	13.70	10.50	10.22	9.99	6.81
	Standard deviation	3.07	3.11	3.17	3.68	4.15	3.70	3.24	3.20	3.16	2.61

Recall	Best	84.71	91.33	92.04	93.96	93.96	96.47	97.34	97.45	97.51	97.87
	Mean	82.62	85.25	87.47	88.54	89.36	90.61	91.25	91.62	93.62	94.71
	Variance	1.98	14.53	16.85	17.50	15.99	20.95	20.62	17.64	9.58	6.74
	Standard deviation	1.41	3.81	4.10	4.18	4.00	4.58	4.54	4.20	3.09	2.60

F1-score	Best	87.73	90.61	91.35	92.75	93.31	94.24	95.48	96.31	96.53	96.63
	Mean	84.76	85.81	86.73	87.59	88.29	88.96	89.73	90.61	91.09	93.16
	Variance	4.37	10.01	12.19	15.03	12.07	13.60	14.83	15.50	13.60	9.17
	Standard deviation	2.09	3.16	3.49	3.88	3.47	3.69	3.85	3.94	3.69	3.03

BT dataset-based statistical analysis

Methods/metrics		SSA-DTCBiNet	Dense-CNN	BA-MCBMM	EDN-SVM	2D CNN-CAE	CJHBA-DRN	Patch ResNet	DE-MCBM	SPO-MCBM	SEnO-DTCBiNet
Accuracy	Best	95.24	89.32	94.84	91.49	94.71	92.19	92.91	94.97	95.63	96.24
	Mean	90.89	85.44	89.21	86.21	88.51	87.35	87.94	90.36	91.80	93.10
	Variance	16.56	9.66	19.58	14.07	19.31	14.63	14.81	17.29	14.61	8.64
	Standard deviation	4.07	3.11	4.43	3.75	4.39	3.82	3.85	4.16	3.82	2.94

Precision	Best	94.05	89.73	93.59	90.62	93.53	91.05	91.37	93.73	94.60	95.49
	Mean	90.41	85.64	88.61	86.06	88.26	87.03	87.59	90.11	91.44	93.02
	Variance	12.50	10.48	14.22	11.83	14.31	10.83	9.82	12.12	11.31	4.85
	Standard deviation	3.54	3.24	3.77	3.44	3.78	3.29	3.13	3.48	3.36	2.20

Recall	Best	97.63	88.49	97.34	93.23	97.07	94.49	96.00	97.47	97.69	97.72
	Mean	91.86	85.05	90.41	86.49	89.00	88.00	88.64	90.86	92.53	93.27
	Variance	26.47	8.24	33.81	19.70	31.94	24.11	28.37	30.91	22.62	20.79
	Standard deviation	5.14	2.87	5.82	4.44	5.65	4.91	5.33	5.56	4.76	4.56

F1-score	Best	95.81	89.11	95.43	91.91	95.27	92.74	93.63	95.56	96.12	96.59
	Mean	91.12	85.34	89.49	86.27	88.62	87.50	88.09	90.47	91.98	93.12
	Variance	18.71	9.27	22.52	15.29	22.03	16.68	17.59	20.20	16.37	11.21
	Standard deviation	4.33	3.04	4.75	3.91	4.69	4.08	4.19	4.49	4.05	3.35

Hyperparameters of the SEnO-DTCBiNet model

Hyperparameters	Values
Kernel size	(3 × 3)
Pooling	(2,2)
Convolution 2D layers	2
Activation function	ReLU
Learning rate	0.02
Dropout rate	0.5
No. of BiLSTM layers	2
LSTM Units	64
Loss function	Categorical-cross entropy
Optimizer	Adam
Number of epochs	100
Pooling	MaxPooling2D
Metrics	Accuracy
Padding	Same
Stride size	2

BraTS 2018 dataset-based statistical analysis

Methods/metrics		Dense-CNN	EDN-SVM	CJHBA-DRN	PatchRes Net	2D-CNN-CAE	BA-MCBM	DE-MCBM	SSA-DTCBiNet	SPO-MCBM	SEnO-DTCBiNet
Accuracy	Best	89.11	89.81	91.60	92.62	93.17	93.75	94.22	94.59	94.78	95.51
	Mean	84.85	85.64	87.00	87.87	88.95	89.39	89.94	90.23	90.96	92.12
	Variance	7.40	7.53	11.51	12.33	12.91	13.83	15.52	15.21	14.40	11.72
	Standard deviation	2.72	2.74	3.39	3.51	3.59	3.72	3.94	3.90	3.79	3.42

Precision	Best	90.35	91.25	91.85	93.03	93.66	94.04	94.56	94.83	94.95	95.37
	Mean	85.40	86.28	87.42	87.93	89.14	89.58	90.22	90.48	91.17	91.98
	Variance	10.32	10.30	11.23	12.63	13.35	13.60	15.39	14.57	15.85	14.38
	Standard deviation	3.21	3.21	3.35	3.55	3.65	3.69	3.92	3.82	3.98	3.79

Recall	Best	86.64	86.95	91.10	91.81	92.18	93.16	93.54	94.13	94.43	95.80
	Mean	83.76	84.37	86.17	87.73	88.57	89.02	89.36	89.73	90.55	92.40
	Variance	3.09	3.44	12.44	12.27	12.36	14.46	16.18	16.68	11.99	7.68
	Standard deviation	1.76	1.85	3.53	3.50	3.52	3.80	4.02	4.08	3.46	2.77

F1-score	Best	88.46	89.05	91.48	92.42	92.92	93.60	94.05	94.48	94.69	95.58
	Mean	84.57	85.31	86.79	87.83	88.85	89.30	89.79	90.10	90.86	92.19
	Variance	6.06	6.23	11.71	12.23	12.73	13.96	15.64	15.57	13.74	10.58
	Standard deviation	2.46	2.50	3.42	3.50	3.57	3.74	3.95	3.95	3.71	3.25

BraTS 2018 dataset-based T-test analysis

Methods/metrics	T-test analysis

	Precision		Accuracy		F1-score		Recall

	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic
SSA-DTCBiNet	0.07	2.73	0.07	2.72	0.07	2.72	0.07	2.69
EDN-SVM	0.13	2.08	0.13	2.09	0.13	2.10	0.13	2.10
CJHBA-DRN	0.10	2.37	0.11	2.23	0.12	2.15	0.15	1.94
PatchResNet	0.11	2.23	0.11	2.29	0.10	2.31	0.10	2.37
Dense-CNN	0.17	1.80	0.17	1.78	0.18	1.77	0.19	1.68
BA-MCBM	0.08	2.66	0.08	2.67	0.08	2.67	0.08	2.67
DE-MCBM	0.07	2.78	0.07	2.75	0.07	2.73	0.08	2.66
2D-CNN-CAE	0.08	2.64	0.08	2.68	0.07	2.69	0.07	2.72
SPO-MCBM	0.07	2.78	0.07	2.72	0.08	2.68	0.08	2.54
SEnO-DTCBiNet	0.07	2.72	0.07	2.73	0.07	2.72	0.08	2.66

Real-time dataset-based T-test analysis

Methods/metrics	T-test analysis

	Precision		F1-score		Accuracy		Recall

	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic
PatchResNet	0.10	2.30	0.10	2.35	0.10	2.34	0.10	2.39
EDN-SVM	0.11	2.25	0.14	2.03	0.13	2.10	0.17	1.83
CJHBA-DRN	0.11	2.29	0.09	2.51	0.09	2.45	0.08	2.63
SPO-MCBM	0.09	2.44	0.10	2.39	0.10	2.41	0.10	2.34
2D-CNN-CAE	0.10	2.35	0.10	2.34	0.10	2.34	0.10	2.32
BA-MCBM	0.10	2.32	0.11	2.25	0.11	2.27	0.12	2.19
DE-MCBM	0.11	2.29	0.12	2.20	0.11	2.23	0.12	2.12
SSA-DTCBiNet	0.09	2.44	0.12	2.18	0.11	2.26	0.14	1.96
Dense-CNN	0.13	2.04	0.14	1.97	0.14	2.00	0.17	1.79
SEnO-DTCBiNet	0.12	2.13	0.12	2.17	0.12	2.16	0.11	2.22

Features based on statistical deep flow

Features	Description	Mathematical notation	Output size
Mean	It is the ratio between the total intensity of all pixels and the total number of pixels within the deep flow feature image.	T1=1r∑i=1rφi {T_1} = {1 \over r}\sum\limits_{i = 1}^r {{\varphi _i}}	[N,120,120,1]
Mean		where φ_i is the feature and r the total images.	[N,120,120,1]
Kurtosis	Kurtosis is defined as the shape of the selected images taken for the statistical measurement.	T2=∑T1−T1¯4rT3 {T_2} = {{\sum {{{\left( {{T_1} - \overline {{T_1}} } \right)}^4}} } \over {r\left( {{T_3}} \right)}}	[N,120,120,1]
Kurtosis		T₂ is the kurtosis	[N,120,120,1]
Standard deviation	The standard deviation is defined as the square root of the variance and represents the average deviation of each pixel intensity from the mean.	T3=1r∑i=1rφi−T12 {T_3} = \sqrt {{1 \over r}\sum\limits_{i = 1}^r {{{\left( {{\varphi _i} - {T_1}} \right)}^2}} }	[N,120,120,1]
Standard deviation		standard deviation is denoted as T₃	[N,120,120,1]
Skew	Skewness is defined as a measure of symmetry of the image.	T4=1r∑i=1rφi−T1T3 {T_4} = {1 \over r}\sum\limits_{i = 1}^r {\left[ {{{{\varphi _i} - {T_1}} \over {{T_3}}}} \right]}	[N,120,120,1]
Skew	Skewness is defined as a measure of symmetry of the image.	T₄ Is the skew	[N,120,120,1]
Variance	Variance is defined as the square of the standard deviation.	T5=T32=1r∑i=1rφi−T12 {T_5} = {\left( {{T_3}} \right)^2} = {1 \over r}\sum\limits_{i = 1}^r {{{\left( {{\varphi _i} - {T_1}} \right)}^2}}	[N,120,120,1]
Variance		T₅ Is the variance	[N,120,120,1]

T-test Analysis based on the BT dataset

Methods/metrics	T-test analysis

	Precision		F1-score		Accuracy		Recall

	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic	p-value	T-statistic
DE-MCBM	0.09	2.54	0.10	2.36	0.09	2.41	0.11	2.24
EDN-SVM	0.14	1.98	0.14	1.96	0.14	1.97	0.15	1.92
CJHBA-DRN	0.10	2.37	0.11	2.25	0.11	2.29	0.12	2.17
PatchResNet	0.09	2.42	0.11	2.22	0.11	2.28	0.13	2.08
Dense-CNN	0.13	2.05	0.12	2.20	0.12	2.15	0.10	2.35
BA-MCBM	0.10	2.38	0.11	2.25	0.11	2.29	0.12	2.15
2D-CNN-CAE	0.10	2.31	0.12	2.16	0.12	2.20	0.13	2.04
SSA-DTCBiNet	0.08	2.62	0.08	2.56	0.08	2.58	0.09	2.51
SPO-MCBM	0.07	2.77	0.07	2.71	0.07	2.73	0.08	2.66
SEnO-DTCBiNet	0.10	2.31	0.07	2.70	0.08	2.62	0.07	2.83

Features based on GLCM

Features	Overview	Formula	Dimensions of outputs
Homogeneity	Homogeneity calculates the similarity of the texture in the distributed gray-level object pairs.	E3=∑kln−1Mkl1+k−l {E_3} = \sum\limits_{kl}^{n - 1} {{{{M_{kl}}} \over {1 + \left\| {k - l} \right\|}}}	[N,120,120,1]
Homogeneity		homogeneity feature is represented as E₃.	[N,120,120,1]
Energy	Energy is used to calculate the uniformity of an image.	E1=∑kln−1Mkl2 {E_1} = \sum\limits_{kl}^{n - 1} {{{\left( {{M_{kl}}} \right)}^2}}	[N,120,120,1]
Energy	Energy is used to calculate the uniformity of an image.	E₁ It is the energy feature, and GLCM of the image Q is denoted as M_kl.	[N,120,120,1]
Entropy	Entropy reproduces the complexity of an image present in the GLCM features.	E4=∑kln−1Mkllog2Mkl {E_4} = \sum\limits_{kl}^{n - 1} {{M_{kl}}{{\log }_2}\left( {{M_{kl}}} \right)}	[N,120,120,1]
Entropy		E₄ Represents the entropy.	[N,120,120,1]
Contrast	Contrast calculates the local variation amounts present in the image.	E5=∑kln−1M(k−l)2 {E_5} = \sum\limits_{kl}^{n - 1} {M{{(k - l)}^2}}	[N,120,120,1]
Contrast		E₅ Depicts the contrast.	[N,120,120,1]
Dissimilarity	It measures the gaps between the mean variances and ROI in the gray-scale image	E2=∑kln−1Mklk−l {E_2} = \sum\limits_{kl}^{n - 1} {{M_{kl}}\left\| {k - l} \right\|}	[N,120,120,1]
Dissimilarity		E₂ Denotes the dissimilarity feature of GLCM.	[N,120,120,1]

Results of the SEnO-DTCBiNet model and comparative methods based on K-fold

Methods/metrics		2D-CNN-CAE	DE-MCBM	Dense-CNN	CJHBA-DRN	Patch ResNet	EDN-SVM	BA-MCBM	SSA-DTCBiNet	SPO-MCBM	SEnO-DTCBiNet
BraTS 2020 dataset	Precision (%)	92.50	93.65	87.57	89.81	92.24	89.33	93.22	95.11	95.39	95.53
	Recall (%)	94.14	97.38	87.89	92.94	93.26	91.92	95.28	97.55	97.69	97.75
	Accuracy (%)	93.05	94.89	87.68	90.85	92.58	90.19	93.91	95.92	96.16	96.27
	F1-Score (%)	93.31	95.48	87.73	91.35	92.75	90.61	94.24	96.31	96.53	96.63

BT dataset	Precision (%)	93.53	93.73	89.73	91.05	91.37	90.62	93.59	94.05	94.60	95.49
	Recall (%)	97.07	97.47	88.49	94.49	96.00	93.23	97.34	97.63	97.69	97.72
	Accuracy (%)	94.71	94.97	89.32	92.19	92.91	91.49	94.84	95.24	95.63	96.24
	F1-Score (%)	95.27	95.56	89.11	92.74	93.63	91.91	95.43	95.81	96.12	96.59

BraTS 2018 dataset	Precision (%)	93.66	94.56	90.35	91.85	93.03	91.25	94.04	94.83	94.95	95.37
	Recall (%)	92.18	93.54	86.64	91.10	91.81	86.95	93.16	94.13	94.43	95.80
	Accuracy (%)	93.17	94.22	89.11	91.60	92.62	89.81	93.75	94.59	94.78	95.51
	F1-Score (%)	92.92	94.05	88.46	91.48	92.42	89.05	93.60	94.48	94.69	95.58

Real-time dataset	Precision (%)	93.50	94.39	89.65	90.83	92.41	90.03	93.54	94.55	94.61	95.78
	Recall (%)	93.96	97.34	84.71	92.04	93.96	91.33	96.47	97.45	97.51	97.87
	Accuracy (%)	93.66	95.38	88.00	91.23	92.92	90.46	94.52	95.52	95.58	96.48
	F1-Score (%)	93.73	95.84	87.11	91.43	93.17	90.67	94.99	95.98	96.04	96.82

BraTS 2020 dataset-based statistical analysis

Methods/metrics		PatchRes Net	BA-MCBM	Dense-CNN	CJHBA-DRN	EDN-DRN	2D-CNN-CAE	SPO-MCBM	SSA-DTCBiNet	DE-MCBM	SEnO-DTCBiNet
Accuracy	Best	92.58	93.91	87.68	90.85	90.19	93.05	94.89	95.92	96.16	96.27
	Mean	87.40	88.59	84.81	86.49	85.78	87.96	89.33	90.24	90.70	92.34
	Variance	14.65	13.72	4.27	10.55	9.03	12.63	14.05	15.28	13.50	10.02
	Standard deviation	3.83	3.70	2.07	3.25	3.01	3.55	3.75	3.91	3.67	3.16

Precision	Best	92.24	93.22	87.57	89.81	89.33	92.50	93.65	95.11	95.39	95.53
	Mean	87.02	87.80	84.92	85.98	85.69	87.28	88.50	89.46	89.89	90.51
	Variance	13.94	14.22	4.09	7.56	7.31	14.04	12.45	14.85	13.31	12.81
	Standard deviation	3.73	3.77	2.02	2.75	2.70	3.75	3.53	3.85	3.65	3.58

Recall	Best	93.26	95.28	87.89	92.94	91.92	94.14	97.38	97.55	97.69	97.75
	Mean	88.17	90.17	84.59	87.50	85.94	89.34	91.00	91.78	92.31	95.99
	Variance	16.34	13.51	4.76	18.41	13.76	10.43	17.68	16.26	13.95	7.93
	Standard deviation	4.04	3.68	2.18	4.29	3.71	3.23	4.20	4.03	3.74	2.82

F1-score	Best	92.75	94.24	87.73	91.35	90.61	93.31	95.48	96.31	96.53	96.63
	Mean	87.59	88.96	84.76	86.73	85.81	88.29	89.73	90.61	91.09	93.16
	Variance	15.03	13.60	4.37	12.19	10.01	12.07	14.83	15.50	13.60	9.17
	Standard deviation	3.88	3.69	2.09	3.49	3.16	3.47	3.85	3.94	3.69	3.03

Tetra-ConvBiNet: MRI-based brain tumor classification using distributed tetra head attention-enabled convolutional bidirectional network

Figures & Tables

Figure 1:

Figure 2:

Figure 3:

Figure 4:

Figure 5:

Figure 6:

Figure 7:

Figure 8:

Figure 9:

Figure 10:

Figure 11:

Figure 12:

Figure 13:

Figure 14:

Figure 15:

Figure 16:

Figure 17:

Figure 18:

Figure 19:

Figure 20:

Figure 21:

Figure 22:

Figure 23:

Figure 24:

Figure 25:

Figure 26:

Figure 27:

Figure 28:

Figure 29:

Figure 30:

T-test analysis based on the BraTS 2020 dataset

Results of the SEnO-DTCBiNet and baseline models based on training percentage

Real-time dataset-based statistical analysis

BT dataset-based statistical analysis

Hyperparameters of the SEnO-DTCBiNet model

BraTS 2018 dataset-based statistical analysis

BraTS 2018 dataset-based T-test analysis

Real-time dataset-based T-test analysis

Features based on statistical deep flow

T-test Analysis based on the BT dataset

Features based on GLCM

Results of the SEnO-DTCBiNet model and comparative methods based on K-fold

BraTS 2020 dataset-based statistical analysis