Auto-Encoder Based Image Classification Technique for Classifying Brain Tumors

Brain tumor diagnosis relies heavily on analyzing MRI images, with computational image analysis techniques playing a crucial role in improving diagnostic accuracy. These techniques aim to classify brain tumors, such as Meningioma, Glioma, and Pituitary tumors, based on their locations, facilitating targeted treatment. A recent neural model-based image classification technique has shown promising results, surpassing non-neural model-based methods. However, it demands a large training set for optimal performance, which can be challenging to obtain. This paper addresses this challenge by proposing a Recurrent Neural Network (RNN) based image classification approach that achieves high accuracy with a smaller training set. The proposed method utilizes a three-layer autoencoder to extract features from high-dimensional image data, which are then classified using a Long Short-Term Memory (LSTM) network. Empirical analysis demonstrates that the proposed technique outperforms contemporary methods in terms of classification accuracy and training set size requirements. The findings suggest that the proposed approach offers a more efficient and accurate method for brain tumor classification, potentially aiding medical professionals in making more informed decisions.