Classification of acute myeloid leukemia by pre-trained deep neural networks: A comparison with different activation functions

• Investigates the efficacy of various activation functions, including Tanh, Softplus, Softsign, ReLU, Leaky ReLU, PreLU, ELU, GELU, SELU, Swish, and Mish, in differentiating WHO classification-based Acute Myeloid Leukemia subtypes. • Classifies AML without maturation, Acute monoblastic leukemia, and Pure erythroid leukemia using VGG16, Inception v3, and ResNet50v2. • GELU activation function performs the best across various architectures in accuracy and efficiency. • ResNet50v2 consistently demonstrates the shortest training times among the three models, regardless of the activation function used. Acute Myeloid Leukemia is a rapidly progressing cancer affecting blood and bone marrow, marked by the swift proliferation of abnormal myeloid cells. Effective treatment requires precise classification of AML subtypes. Conventional classification methods rely on manual microscopic analysis, which is time-consuming and variable, while traditional machine learning approaches often struggle with feature extraction and generalization. This underscores the need for automated detection using deep learning models, particularly CNN architectures, which enhance this process by minimizing manual effort. Activation functions are critical in neural networks, introducing non-linearity that influences training convergence, computational efficiency, and model performance. This study evaluates the effectiveness of CNN architectures, Sequential (VGG16), Directed Acyclic Graph (InceptionV3), and Residual (ResNet50v2) in distinguishing between AML subtypes: AML without maturation, Acute Monoblastic Leukemia, and Pure Erythroid Leukemia, using peripheral blood smear images, while also investigating the impact of different activation functions on model accuracy and training time. The results show that ResNet50v2 achieves the shortest training time, while InceptionV3 takes the longest due to its complex architecture. GELU delivers the highest accuracy, reaching 94.02% in InceptionV3 and 92.53% in ResNet50v2, while SELU achieves the highest accuracy for VGG16 at 92.83%. Mish provides competitive accuracy with lower training time than GELU, while Softplus and Softsign consistently perform poorly. This research demonstrates the potential of CNNs for automating AML subtype classification and identifies GELU as the most effective activation function. Future work could explore data augmentation, optimized activation functions, and attention mechanisms to improve classification performance.