Violence recognition is challenging since recognition must be performed on videos acquired by a lot of surveillance cameras at any time or place. It should make reliable detections in real time and inform surveillance personnel promptly when violent crimes take place. Therefore, we focus on efficient violence recognition for real-time and on-device operation, for easy expansion into a surveillance system with numerous cameras. In this paper, we propose a novel violence detection pipeline that can be combined with the conventional 2-dimensional Convolutional Neural Networks (2D CNNs). In particular, frame-grouping is proposed to give the 2D CNNs the ability to learn spatio-temporal representations in videos. It is a simple processing method to average the channels of input frames and group three consecutive channel-averaged frames as an input of the 2D CNNs. Furthermore, we present spatial and temporal attention modules that are lightweight but consistently improve the performance of violence recognition. The spatial attention module named Motion Saliency Map (MSM) can capture salient regions of feature maps derived from the motion boundaries using the difference between consecutive frames. The temporal attention module called Temporal Squeeze-and-Excitation (T-SE) block can inherently highlight the time periods that are correlated with a target event. Our proposed pipeline brings significant performance improvements compared to the 2D CNNs followed by the Long Short-Term Memory (LSTM) and much less computational complexity than existing 3D-CNN-based methods. In particular, MobileNetV3 and EfficientNet-B0 with our proposed modules achieved state-of-the-art performance on six different violence datasets. Our codes are available at https://github.com/ahstarwab/Violence_Detection.

We propose approaches to the problem of multi-paths beam tracking in millimeter wave (mmWave) mobile communications. Precise angle tracking is necessary to take advantage of beamforming technology in this mmWave band. The mobile channel is a highly nonlinear function with respect to the state that comprises the beam angle and the channel path gain. Therefore, the beam angle tracking is challenging and crucial. We propose a couple of particle filtering frameworks for beam tracking, and show outperforming result of auxiliary particle filtering approach over recently proposed tracking approaches.

Although sound source localization is a desirable technique in many communication systems and intelligence applications, the distortion caused by diffuse noise or reverberation makes the time delay estimation (TDE) between signals acquired by a pair of microphones a complicated and challenging problem. In this paper, we describe a method that can efficiently achieve sound source localization in noisy and reverberant environments. This method is based on the generalized cross-correlation (GCC) function with phase transform (PHAT) weights (GCC-PHAT) to achieve robustness against reverberation. In addition, to estimate the time delay robust to diffuse components and to further improve the robustness of the GCC-PHAT against reverberation, time-frequency(t-f) components of observations directly emitted by a point source are chosen by “inversed” diffuseness. The diffuseness that can be estimated from the coherent-to-diffuse power ratio (CDR) based on spatial coherence between two microphones represents the contribution of diffuse components on a scale of zero to one with direct sounds from a source modeled to be fully coherent. In particular, the “inversed” diffuseness is binarized with a very rigorous threshold to select highly reliable components for accurate TDE even in noisy and reverberant environments. Experimental results for both simulated and real-recorded data consistently demonstrated the robustness of the presented method against diffuse noise and reverberation.

In this paper, we propose a new approach for tracking a target maneuvering based on the multiple-constant-turns model. Usually, the interactive-multiple-model (IMM) algorithm based on the extended Kalman filter (EKF) is employed for this problem with successful tracking performance. Recently proposed IMM-particle filter (PF) shows outperforming results for this nonlinear problem. The proposed approach is a new framework of particle filtering (PF) that adopts the minimax strategy to standard PF. The minimax strategy results in the decreased variance of the weights of particles that provides the robustness against degeneracy phenomenon that is a common problem of generic particle filtering. In this paper, we show outperforming results by IMM-MPF over IMM-PF besides the IMM-EKF in terms of various aspects.

In this article, we propose a method to monitor traffic on the road using a 24.1 GHz radar system. In our work, we design our own traffic monitoring radar system from radar hardware, such as antenna systems and the transceiver, to radar software, such as detection and tracking algorithms. The radar we developed uses frequency-modulated continuous waveforms with a center frequency of 24.1 GHz, and its range resolution is tens of centimeters, which is enough to distinguish vehicles in multiple lanes. In addition, we can effectively distinguish vehicles by estimating their angles using an array antenna system. To evaluate the performance of our radar, we install it on the pedestrian overpass facing the road. From actual measurement results, our radar system can reliably estimate the distance to the vehicle, the speed of the vehicle, and the angle of the vehicle for a total of four lanes up to 300 m in real time. The traffic monitoring system using a radar sensor can compensate for the disadvantages of the camera-based road monitoring method.