Anemia Identification from Blood Smear Images Using Deep Learning: An XAI Approach

Anemia is a prevalent hematologic disorder characterized by a reduction in the number of red blood cells or a decrease in their oxygen-carrying capacity. Accurate and timely diagnosis of anemia is crucial for effective patient management and improved health outcomes. In this proposed work, we developed a CNN-based model to classify individuals as having symptoms of Iron-deficiency Anemia (IDA) or not, using blood smear images. The CNN model will be trained and fine-tuned using this dataset to maximize the classification accuracy. This helps healthcare professionals to understand the system's decisions and assist in the diagnosis and management of anemia. Despite a relatively limited number of data samples, our proposed work showed promising results by leveraging augmentation techniques. We achieved an accuracy of 0.95, with precision and recall values of 0.9 and 0.89, respectively. The area under the Receiver Operating Characteristic (ROC) curve is reported as 0.98.

I. Introduction

Anemia is a medical condition characterized by a deficiency of healthy red blood cells, leading to inadequate oxygen supply to the body's tissues. It is often associated with low hemoglobin levels and can cause fatigue and weakness. Anemia can manifest in various forms, each with its own underlying cause, and its severity can range from mild to severe. It is important to seek medical attention if you suspect you have anemia, as it can be an indication of a more serious underlying illness.