Dr Nellutla Sasikala

https://doi.org/10.64823/ijter.2504008

For computer vision tasks like object detection, recognition, and classification, relay on feature extraction, labelling and segmentation of captured videos or images. Applications like smart city, health care, geoscience and remote sensing are based on video analysis. Image segmentation for video analysis plays an vital role. One of the novel segmentation strategies which has been recently developed is panoptic segmentation. Panoptic segmentation is a fusion of semantic and instance segmentation. In self autonomous driving, medical image analysis, crowd counting, etc. have complicated background components, the high variability of object appearances, numerous overlapping objects and ambiguous object boundaries makes the task challenging. For such applications panoptic segmentation is used which provides several state-of-art methods and robust learning.

Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) have evolved from rule-based symbolic systems into data-driven computational methods capable of perception, prediction, decision support, and content generation. This chapter presents an accessible yet technically grounded overview of the relationship among AI, ML, and DL, tracing their historical development from early symbolic reasoning and theoretical foundations to modern neural networks and transformer-based systems. It explains the principal ML paradigms supervised, unsupervised, reinforcement, semi-supervised, and self-supervised learning and introduces widely used algorithms including regression, decision trees, ensemble methods, support vector machines, k-nearest neighbors, Naive Bayes, gradient boosting, and k-means clustering. The chapter then examines deep-learning architectures such as Convolutional neural networks, recurrent and long short-term memory networks, transformers, generative adversarial networks, and diffusion models, together with training concepts including back propagation, gradient descent, regularization, transfer learning, and evaluation metrics. Applications across healthcare, finance, transportation, manufacturing, agriculture, education, cyber security, and creative work are discussed, alongside challenges involving bias, explainability, privacy, misinformation, employment, and computational and environmental costs. The chapter concludes by emphasizing human-centered deployment, responsible governance, interpretability, and continuous evaluation as AI systems become increasingly integrated into high-impact domains.