EXPLAINABLE AI FRAMEWORK FOR EARLY ORAL CANCERPREDICTION USING DEEP LEARNING

Oral cancer is a major heath concern that effects the thousands of people every year, early detection plays a crucial role in improving treatment outcomes and increasing the chances of survival. Deep learning has shown promise in analysing medical images automatically. Many of these models work like black boxes. This means it's hard to understand how they make decisions. That lack of clarity can be a problem in healthcare, where trust and transparency essential. This study introduces an artificial intelligence framework for predicting oral cancer early. The goal is to help doctors by providing insights into predictions. The system uses learning and combines information from different sources. It looks at both pictures of the mouth and clinical data. The clinical data includes things like smoking history, use of tobacco and alcohol hemoglobin levels, fever, weight loss trouble swallowing, white or red patches and non-healing sores. Before any prediction is made there is a validation step. It helps avoid errors caused by unrelated inputs. An EfficientNet model is used to extract features from the images. At the time the structured clinical data is processed separately and then combined with the image features. This fusion allows for predictions based on both visual and patient-specific details. Grad- CAM highlights areas in the images. LIME explains predictions by approximating the model locally. SHAP provides explanations that show which features contribute most to the outcome. These tools together give a complete picture of why a prediction was made.

Authors

Publication Details

Journal
Scientiarum: A multidisciplinary journal
Published
2026-11-12
DOI
https://doi.org/10.54646/tamxqb35
Primary Topic
Explainable Artificial Intelligence (XAI)
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

EXPLAINABLE AI FRAMEWORK FOR EARLY ORAL CANCERPREDICTION USING DEEP LEARNING

Parnapalli Pushpanjali, Navya R
Scientiarum: A multidisciplinary journal
Explainable Artificial Intelligence (XAI)
article

EXPLAINABLE AI FRAMEWORK FOR EARLY ORAL CANCERPREDICTION USING DEEP LEARNING

Parnapalli Pushpanjali, Navya R
article en

Abstract

Oral cancer is a major heath concern that effects the thousands of people every year, early detection plays a crucial role in improving treatment outcomes and increasing the chances of survival. Deep learning has shown promise in analysing medical images automatically. Many of these models work like black boxes. This means it's hard to understand how they make decisions. That lack of clarity can be a problem in healthcare, where trust and transparency essential. This study introduces an artificial intelligence framework for predicting oral cancer early. The goal is to help doctors by providing insights into predictions. The system uses learning and combines information from different sources. It looks at both pictures of the mouth and clinical data. The clinical data includes things like smoking history, use of tobacco and alcohol hemoglobin levels, fever, weight loss trouble swallowing, white or red patches and non-healing sores. Before any prediction is made there is a validation step. It helps avoid errors caused by unrelated inputs. An EfficientNet model is used to extract features from the images. At the time the structured clinical data is processed separately and then combined with the image features. This fusion allows for predictions based on both visual and patient-specific details. Grad- CAM highlights areas in the images. LIME explains predictions by approximating the model locally. SHAP provides explanations that show which features contribute most to the outcome. These tools together give a complete picture of why a prediction was made.

Scientiarum: A multidisciplinary journal
Explainable Artificial Intelligence (XAI)
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon