Big data analytics (BDA) is a systematic approach for analyzing and identifying different patterns, relations, and trends within a large volume of data. In this paper, we apply BDA to criminal data where exploratory data analysis is conducted for visualization and trends prediction. Several the state-of-the-art data mining and deep learning techniques are used. Following statistical analysis and visualization, some interesting facts and patterns are discovered from criminal data in San Francisco, Chicago, and Philadelphia. The predictive results show that the Prophet model and Keras stateful LSTM perform better than neural network models, where the optimal size of the training data is found to be three years. These promising outcomes will benefit for police departments and law enforcement organizations to better understand crime issues and provide insights that will enable them to track activities, predict the likelihood of incidents, effectively deploy resources and optimize the decision making process.

Identifying tiny objects from extremely low resolution (LR) UAV-based remote sensing images is generally considered as a very challenging task, because of very limited information in the object areas. In recent years, there have been very limited attempts to approach this problem. These attempts intend to deal with LR image classification by enhancing either the poor image quality or image representations.In this paper, we argue that the performance improvement in LR image classification is affected by the inconsistency of the information loss and learning priority on Low-Frequency (LF) components and High-Frequency (HF) components.To address this LF-HF inconsistency problem, we propose a Dual-Stream Representation Learning Generative Adversarial Network (DRL-GAN).The core idea is to produce super image representations optimal for LR recognition by simultaneously recovering the missing information in LF and HF components, respectively, under the guidance of high-resolution (HR) images.We evaluate the performance of DRL-GAN on the challenging task of LR image classification.A comparison of the experimental results on the LR benchmark, namely HRSC and CIFAR-10, and our newly collected “WIDER-SHIP” dataset demonstrates the effectiveness of our DRL-GAN, which significantly improves the classification performance, with up to 10% gain on average.

Identifying tiny objects with extremely low resolution is generally considered a very challenging task even for human vision, due to limited information presented inside the object areas. There have been very limited attempts in recent years to deal with low-resolution recognition. The existing solutions rely on either generating super-resolution images or learning multi-scale features. However, their performance improvement becomes very limited, especially when the resolution becomes very low. In this paper, we propose a Representation Learning Generative Adversarial Network (RL-GAN) to generate super image representation that is optimized for recognition. Our solution deals with the classical vision task of object recognition in the distance. We evaluate our idea on the challenging task of low-resolution object recognition. Comparison of experimental results conducted on public and our newly created WIDER-SHIP datasets demonstrate the effectiveness of our RL-GAN, which improves the classification results significantly, with 10-15% gain on average, compared with benchmark solutions.

Aiming at the problem that micro-expression (ME) participants have their own expressions that interfere with microexpression recognition (MER), a dual-stream high-dimensional features implicit augmentation for MER is proposed. The innovative fusion method of adaptive face mask and Action Units (AUs) is used to reduce the interference of the participants' own expressions, and combined with optical flow information to form more focused high-dimensional spatiotemporal information. Enhance high-dimensional spatio-temporal features through an improved implicit semantic data enhancement algorithm to alleviate the problem of small sample data. Using high-dimensional spatiotemporal information to focus the network model on ME active areas. Using a two-stream hybrid convolutional network to extract features from temporal and spatial domain information respectively to improve the representation ability of the network. The experimental results show that under the evaluation criteria of MEGC2019, UF1 and UAR are improved by 5.5% and 6.8% respectively compared with the best method.