• Vankayalapati Sravani
• Panduri Baby Sri Karthikeyi
• Uppe Jhansi Lakshmi
• Penmetsa Bhavya Jaya Sree

The advent of deep learning has revolutionized medical image analysis, offering promising avenues for early disease detection and diagnosis. This paper presents a novel approach utilizing DenseNet models for the detection of leukemia, a critical hematologic malignancy with diverse manifestations across different patient populations. DenseNet, characterized by its dense connections between layers, has demonstrated superior performance in various image classification tasks by enhancing feature propagation and encouraging feature reuse. Leveraging this architecture, we propose a robust framework for automatic leukemia detection from peripheral blood smear images. The proposed model first preprocesses the input images to enhance contrast and reduce noise, thereby improving the quality of feature extraction. Subsequently, DenseNet is employed to extract hierarchical representations of the input images, capturing intricate patterns and structures indicative of leukemia presence. Transfer learning techniques are employed to adapt pre-trained DenseNet models to the specific task of leukemia detection, leveraging knowledge learned from large-scale datasets. To evaluate the effectiveness of the proposed approach, extensive experiments are conducted on publicly available leukemia image datasets. Quantitative metrics such as accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC-ROC) are employed to assess the model's performance. Moreover, qualitative analysis is performed to interpret the model's decision-making process and identify salient regions contributing to classification. Results demonstrate the efficacy of the proposed DenseNet-based framework in accurately detecting leukemia from peripheral blood smear images, outperforming traditional machine learning approaches and competing deep learning architectures. The proposed model exhibits high sensitivity and specificity, crucial for reliable disease diagnosis in clinical settings. Overall, this study showcases the potential of DenseNet models in enhancing leukemia detection accuracy, thereby aiding clinicians in timely and accurate diagnosis, ultimately improving patient outcomes and healthcare delivery.

DenseNet, RandomSearch, Matplotlib, LabelEncoder, NumPy, Sklearn, Keras, TensorFlow.

"Lukemia Detection Using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 4, page no.h226-h230, April-2024, Available :http://www.jetir.org/papers/JETIR2404728.pdf

"Lukemia Detection Using Deep Learning", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 4, page no. pph226-h230, April-2024, Available at : http://www.jetir.org/papers/JETIR2404728.pdf