Exploiting Multispectral And Contextual Information To Improve Human Detection

Exploiting Multispectral and Contextual Information to Improve Human Detection

Human detection has various applications, e.g., autonomous driving car, surveillance system, and retail. In this dissertation, we first exploit multispectral images (i.e., RGB and thermal images) for human detection. We extensively analyze Faster R-CNN for the detection task and then model multispectral human detection into a fusion problem of convolutional networks (ConvNets). We design four distinct ConvNet fusion architectures that integrate two-branch ConvNets on different stages of neural networks, all of which yield better performance compared with the baseline detector. In the second part of this dissertation, we leverage instance-level contextual information in crowded scenes to boost performance of human detection. Based on a context graph that incorporates both geometric and social contextual patterns from crowds, we apply progressive potential propagation algorithm to discover weak detections that are contextually compatible with true detections while suppressing irrelevant false alarms. The method significantly improves the performance of any shallow human detectors, obtaining comparable results to deep learning based methods.

Image Understanding and the Man-machine Interface III

Synthetic Aperture Radar (SAR) Data Applications

This carefully curated volume presents an in-depth, state-of-the-art discussion on many applications of Synthetic Aperture Radar (SAR). Integrating interdisciplinary sciences, the book features novel ideas, quantitative methods, and research results, promising to advance computational practices and technologies within the academic and industrial communities. SAR applications employ diverse and often complex computational methods rooted in machine learning, estimation, statistical learning, inversion models, and empirical models. Current and emerging applications of SAR data for earth observation, object detection and recognition, change detection, navigation, and interference mitigation are highlighted. Cutting edge methods, with particular emphasis on machine learning, are included. Contemporary deep learning models in object detection and recognition in SAR imagery with corresponding feature extraction and training schemes are considered. State-of-the-art neural network architectures in SAR-aided navigation are compared and discussed further. Advanced empirical and machine learning models in retrieving land and ocean information — wind, wave, soil conditions, among others, are also included.