Applied Computer Vision through Artificial Intelligence

9Understanding the Unseen: Explainability in Deep Learning for Computer Vision

Apoorva Jain¹*, Jasminder Kaur Sandhu¹ and Pulkit Dwivedi²

¹School of Computer Science and Engineering, IILM University, Greater Noida, India

²School of Computer Science Engineering & Technology, Bennett University, Greater Noida, India

Abstract

This book chapter, Explainability in Deep Learning for Computer Vision, explores the critical role of interpretability in the application and adoption of deep learning models across various real-world sectors, including healthcare, finance, and autonomous systems. The chapter begins with an overview of the transformative impact of deep learning in computer vision, highlighting significant advancements in image classification, object detection, and semantic segmentation. It underscores the capability of Convolutional Neural Networks (CNNs) to automatically extract relevant features from extensive visual data, surpassing traditional manual feature engineering methods. As deep learning models grow increasingly complex and are deployed in mission-critical applications, interpretability and explainability emerge as paramount considerations. These attributes enable users to comprehend model decisions, establish trust, ensure compliance with ethical and regulatory standards, and facilitate debugging and model improvement. The chapter delves into the necessity of interpretability for understanding model behavior, fostering stakeholder trust, and promoting ethical ...