A comprehensive overview of deep learning models for object detection from videos/images

TL;DR

This paper provides a comprehensive review of deep learning techniques for object detection in videos and images, emphasizing architectural innovations, generative models, and temporal information to improve robustness and accuracy.

Contribution

It classifies and evaluates modern deep learning methods based on architectures, data strategies, and surveillance challenges, offering insights into current effectiveness and future trends.

Findings

CNN-based detectors are effective in various scenarios.

Generative models aid in reconstructing missing data and reducing occlusions.

Temporal fusion improves detection robustness in dynamic environments.

Abstract

Object detection in video and image surveillance is a well-established yet rapidly evolving task, strongly influenced by recent deep learning advancements. This review summarises modern techniques by examining architectural innovations, generative model integration, and the use of temporal information to enhance robustness and accuracy. Unlike earlier surveys, it classifies methods based on core architectures, data processing strategies, and surveillance specific challenges such as dynamic environments, occlusions, lighting variations, and real-time requirements. The primary goal is to evaluate the current effectiveness of semantic object detection, while secondary aims include analysing deep learning models and their practical applications. The review covers CNN-based detectors, GAN-assisted approaches, and temporal fusion methods, highlighting how generative models support tasks such as reconstructing missing frames, reducing occlusions, and normalising illumination. It also outlines preprocessing pipelines, feature extraction progress, benchmarking datasets, and comparative evaluations. Finally, emerging trends in low-latency, efficient, and spatiotemporal learning approaches are identified for future research.