Litcius/Paper detail

Survey: Convolution Neural networks in Object Detection

Heba Hakim, Ali Fadhil

2021Journal of Physics Conference Series24 citationsDOIOpen Access PDF

Abstract

Abstract In latest years, deep neural networks were observed to be the most influential among all innovations in the computer vision field, generating remarkable performance on image classification. Convolution neural networks (CNNs) are considering as an interesting tool for studying vision of biological because this category of artificial vision systems shows the capabilities of visual recognition similar to those of human observers. By improving the recognition performance of these models, it appears that they become more effective in prediction. Recent benchmarks have shown that deep CNNs are excellent approaches for object recognition and detection. In this paper, we are focusing on the core building blocks of convolution neural networks architecture. Different object detection methods that utilize convolution neural networks are discussed and compared. On the other hand, there is a simple summary of the common CNNs architectures.

Topics & Concepts

Convolutional neural networkConvolution (computer science)Artificial intelligenceComputer scienceCognitive neuroscience of visual object recognitionArtificial neural networkObject detectionDeep learningObject (grammar)Pattern recognition (psychology)Field (mathematics)ArchitectureComputer visionMathematicsGeographyPure mathematicsArchaeologyAdvanced Neural Network ApplicationsHuman Pose and Action RecognitionAnomaly Detection Techniques and Applications