China
Africa
Explore chapters and articles related to this topic
Deep Learning Solutions for Pest Detection
Published in S. Poonkuntran, Rajesh Kumar Dhanraj, Balamurugan Balusamy, Object Detection with Deep Learning Models, 2023
In recent years, “object detection” has become one of the most challenging problems to receive attention in computer vision applications. Nowadays, images are available everywhere, and the critical content in an image is the object. Extracting the semantic information from images for further analysis like object recognition and detection is a primary image processing application. There have been numerous endeavors to utilize noteworthy features of object detection for different applications like video surveillance, autonomous driving, human-computer interfaces, robotic vision, crowd counting, anomaly detection, and healthcare. The emergence of deep neural networks (DNNs) leads to learning more complex features effectively. The convolutional neural network (CNN) features have the highest representative power compared with traditional approaches. This article investigates object detection in agriculture; in particular, it focuses on how deep learning-based object detection can be used for automatic pest detection in paddy crops.
Deep Learning for Analyzing the Data on Object Detection and Recognition
Published in R. Sujatha, S. L. Aarthy, R. Vettriselvan, Integrating Deep Learning Algorithms to Overcome Challenges in Big Data Analytics, 2021
Object detection refers to locating objects from digital images and classifying them. An issue of rudimentary nature for solving is object detection; existing methods have been developed. Also, in real-time implementations, there exists a wide potential to develop specific mechanisms and object detection as simple services. In a number of applications, CNN-pivoted object detectors are utilized with progressive outcomes from CNNs of the deep type of architectural setups. It is classified as either single-stage or two-stage target detecting model depending on methodology. As far as invariant occlusions, sizes, lighting, intra-class differences, and deformations are concerned, object detection in real-time scenarios using various GPU-pivoted embedded platforms was expected to be strong. For these reasons, the ability to identify small objects in a video is high, which decreases the efficiency of tracking systems for real-time objects.
Pertinent Properties of Euclidean Space
Published in Gerhard X. Ritter, Gonzalo Urcid, Introduction to Lattice Algebra, 2021
Gerhard X. Ritter, Gonzalo Urcid
The data for facial recognition are faces, and the sensors are composed of one or more cameras. A single camera provides a 2-dimensional image while three or more cameras are usually employed for obtaining a 3-dimensional image. Because of business involvement and continued research advancement there exist a wide variety of different types of competing algorithms for face recognition. Different algorithms often require different preprocessing tasks. For instance, some tasks may involve finding faces in a crowd and tracking one or more of these faces. The detection of faces in images is a subset of computer vision (CV) algorithms that deal with object detection in imagery. Object detection is an extremely active area of research, owing to the fact of its applications in such areas as driverless vehicles, national defense, and medical diagnostics. If the aim is face recognition, then face detection is more often than not an essential first step.
Review of CNN in aerial image processing
Published in The Imaging Science Journal, 2023
Xinni Liu, Kamarul Hawari Ghazali, Fengrong Han, Izzeldin Ibrahim Mohamed
Object detection is a crucial research topic in image processing and computer vision, where it focused on detecting a certain class of semantic objects such as buildings, cars or humans in videos and digital images. In recent years, the accuracy of object detection has greatly increased due to the success of deep learning-based CNNs. Currently, deep learning methods used for object detection can be divided into two classes, namely one stage detection and two stage detection [74–77]. The latter involves two steps. First, the regions that potentially contain objects are proposed. Then, the proposed regions obtained in the first step will be classified into different categories. However, the regions proposal step is removed in one stage detection, whereby the object localization and classification are completed at one step, greatly improving the detection speed compared with two stage detection. Examples of well-researched fields for object detection contain pedestrian detection and face detection [39].
A Comparative Assessment of Deep Neural Network Models for Detecting Obstacles in the Real Time Aerial Railway Track Images
Published in Applied Artificial Intelligence, 2022
R. S. Rampriya, R. Suganya, Sabari Nathan, P. Shunmuga Perumal
Accordingly, computer vision (CV) and artificial intelligence-based methods play a significant role in detecting obstacles, especially in Railway Industry. The traditional CV method uses ROI (Region of Interest) for rail extraction and the Sobel edge detection technique for stationary object detection (Ukai 2004). The optical flow method is applied between frames to detect dynamic hazardous obstacles and neglect irrelevant background objects (Uribe et al., 2012). Artificial intelligence enabled improvement in deep neural network technology and excellent advancement in object detection. The Faster R-CNN is used for object detection on the detected rail tracks in which canny edge detection and Hough transform are used for ROI that is rail track (Kapoor, Goel, and Sharma 2018). A multi-level obstacle detection method is presented with two parts: the creation of a feature map using Residual Neural Network (RNN) for object detection of various sizes at different distances followed by a sequence of convolution layers are implemented for feature extraction, which draws bounding boxes and calculates confidence score (Xu et al. 2019). On the other hand, DisNet is presented with two steps: the first, YOLOv3, for object detection, and the second is a multiple hidden layers network for distance estimation (Risti et al. 2020).
A Technique for Human Upper Body Parts Movement Tracking
Published in IETE Journal of Research, 2022
Krishan Kumar, Abhya Mishra, Sanjay Dahiya, Ajay Kumar
In everyday life, images and videos are multiplied exponentially because of the rise of mobile devices. The main feature of mobile is to generate images and videos. Consequently, it has enhanced the image's overall processing, such as image enhancement, Noise reduction, etc. Object detection operates solid ground in the images and videos. Commonly, these grounds approach the damage and crimes to investigate an unauthorized offense, public safety, etc. Detection is often implemented to distinguish between the object present inside images or videos and identify. However, images and videos are now an essential part of human life. Object detection is widely considered an exciting field of research. It has different aspects where it can be used, like detecting the main object from an image. Moreover, it is referred to as salient object detection. Object detection in computer vision and image processing possesses many applications such as face detection and recognition, salient object detection, the total number of people present at the venue, path tracking, vehicle detection, obtaining a particular object through Detection, etc.