When multiple object tracking (MOT) based on the tracking-by-detection paradigm is implemented, the similarity metric between the current detections and existing tracks plays an essential role. Most of the MOT schemes based on a deep neural network (DNN) learn the similarity metric using a Siamese architecture, but the plain Siamese architecture might not be enough owing to its structural simplicity and lack of motion information. This study aims to propose a new MOT scheme to overcome the existing problems in the conventional MOTs. Feature Pyramid Siamese Network (FPSN) is proposed to address the structural simplicity. The FPSN is inspired by a feature pyramid network (FPN) and it extends the Siamese network by applying FPN to the plain Siamese architecture and by developing a new multi-level discriminative feature. A spatiotemporal motion feature is added to the FPSN to overcome the lack of motion information and to enhance the performance in MOT. Thus, FPSN-MOT considers not only the appearance feature but also motion information. Finally, FPSN-MOT is applied to the public MOT Challenge benchmark problems and its performance is compared to that of the other state-of-the-art MOT methods.

Most of the commercial nighttime pedestrian detection (PD) methods reported previously utilized the histogram of oriented gradient (HOG) or the local binary pattern (LBP) as the feature and the support vector machine (SVM) as the classifier using thermal camera images. In this paper, we propose a new feature called the thermal-position-intensity-histogram of oriented gradient (TPIHOG or T $\mathsf{\pi }$ HOG) and developed a new combination of the T $\mathsf{\pi }$ HOG and the additive kernel SVM (AKSVM) for efficient nighttime pedestrian detection. The proposed T $\mathsf{\pi }$ HOG includes detailed information on gradient location; therefore, it has more distinctive power than the HOG. The AKSVM performs better than the linear SVM in terms of detection performance, while it is much faster than other kernel SVMs. The combined T $\mathsf{\pi }$ HOG-AKSVM showed effective nighttime PD performance with fast computational time. The proposed method was experimentally tested with the KAIST pedestrian dataset and showed better performance compared with other conventional methods.

Scene recognition is a kind of image recognition problems which is aimed at predicting the category of the place at which the image is taken. In this paper, a new scene recognition method using the convolutional neural network (CNN) is proposed. The proposed method is based on the fusion of the object and the scene information in the given image and the CNN framework is named as FOS (fusion of object and scene) Net. To combine the object and the scene information effectively, a new fusion framework named CCG (correlative context gating) is proposed. In addition, a new loss named scene coherence loss (SCL) is developed to train the FOSNet and to improve the scene recognition performance. The proposed SCL is based on the idea that the scene class does not change all over the image. The proposed FOSNet was experimented with three most popular scene recognition datasets, and their state-of-the-art performance is obtained in two sets: 60.14% on Places 2 and 90.30% on MIT indoor 67. The second highest performance of 77.28% is obtained on SUN 397.

Although much research has taken place in WiFi indoor localization systems, their accuracy can still be improved. When designing this kind of system, fingerprint-based methods are a common choice. The problem with fingerprint-based methods comes with the need of site surveying the environment, which is effort consuming. In this work, we propose an approach, based on support vector regression, to estimate the received signal strength at non-site-surveyed positions of the environment. Experiments, performed in a real environment, show that the proposed method could be used to improve the resolution of fingerprint-based indoor WiFi localization systems without increasing the site survey effort.

For an autonomous ship to navigate safely and avoid collisions with other ships, reliably detecting and classifying nearby ships under various maritime meteorological environments is essential. In this paper, a novel probabilistic ship detection and classification system based on deep learning is proposed. The proposed method aims to detect and classify nearby ships from a sequence of images. The method considers the confidence of a deep learning detector as a probability; the probabilities from the consecutive images are combined over time by Bayesian fusion. The proposed ship detection system involves three steps. In the first step, ships are detected in each image using Faster region-based convolutional neural network (Faster R-CNN). In the second step, the detected ships are gathered over time and the missed ships are recovered using the Intersection over Union of the bounding boxes between consecutive frames. In the third step, the probabilities from the Faster R-CNN are combined over time and the classes of the ships are determined by Bayesian fusion. To train and evaluate the proposed system, we collected thousands of ship images from Google image search and created our own ship dataset. The proposed method was tested with the collected videos and the mean average precision increased by 89.38 to 93.92% in experimental results.