Название: Advancement of Deep Learning and its Applications in Object Detection and Recognition
Автор: Roohie Naaz Mir, Vipul Kumar Sharma, Ranjeet Kumar Rout
Издательство: River Publishers
Год: 2023
Страниц: 319
Язык: английский
Формат: pdf (true)
Размер: 26.1 MB

Object detection is a basic visual identification problem in computer vision that has been explored extensively over the years. Visual object detection seeks to discover objects of specific target classes in a given image with pinpoint accuracy and apply a class label to each object instance. Object recognition strategies based on Deep Learning have been intensively investigated in recent years as a result of the remarkable success of Deep Learning-based image categorization. In this book, we go through in detail detector architectures, feature learning, proposal generation, sampling strategies, and other issues that affect detection performance.

The book describes every newly proposed novel solution but skips through the fundamentals so that readers can see the field's cutting edge more rapidly. Moreover, unlike prior object detection publications, this project analyses Deep Learning-based object identification methods systematically and exhaustively, and also gives the most recent detection solutions and a collection of noteworthy research trends. The book focuses primarily on step-by-step discussion, an extensive literature review, detailed analysis and discussion, and rigorous experimentation results. Furthermore, a practical approach is displayed and encouraged.

Estimating crowd counting from images is a difficult but important task given the large range of applications such as public safety, traffic monitoring, and urban planning. Occlusions, uneven density, variation in scale and perspective are all challenges in crowd analysis. Thanks to advancements in Deep Learning and constructing demanding databases, modern computer vision techniques have led to numerous cutting-edge methods that build the abilities needed to properly execute a wide range of scenarios. This article presents a brief description of pioneering methods based on hand-crafted representations, followed by an examination of contemporary approaches based on convolutional neural networks (CNNs) that have achieved significant performance. In addition, the most frequently utilized datasets are addressed, and lastly, promising research routes in this rapidly increasing area are indicated.

Object detection serves as a significant step in improving performance of complex downstream computer vision tasks. It has been extensively studied for many years now and current state-of-the-art 2D object detection techniques proffer superlative results even in complex images. In this chapter, we discuss the geometry-based pioneering works in object detection, followed by the recent breakthroughs that employ Deep Learning. Some of these use a monolithic architecture that takes a RGB image as input and passes it to a feed-forward ConvNet or vision Transformer. These methods, thereby predict class-probability and bounding-box coordinates, all in a single unified pipeline. Two-stage architectures on the other hand, first generate region proposals and then feed it to a CNN to extract features and predict object category and bounding-box. We also elaborate upon the applications of object detection in video event recognition, to achieve better fine-grained video classification performance. Further, we highlight recent datasets for 2D object detection both in images and videos, and present a comparative performance summary of various state-of-the-art object detection techniques.