Multiple Object Tracking (MOT) plays an important role in solving many fundamental problems in video analysis and computer vision. Most MOT methods employ two steps: Object Detection and Data Association. The first step detects objects of interest in every frame of a video, and the second establishes correspondence between the detected objects in different frames to obtain their tracks. Object detection has made tremendous progress in the last few years due to deep learning. However, data association for tracking still relies on hand crafted constraints such as appearance, motion, spatial proximity, grouping etc. to compute affinities between the objects in different frames. In this paper, we harness the power of deep learning for data association in tracking by jointly modeling object appearances and their affinities between different frames in an end-to-end fashion. The proposed Deep Affinity Network (DAN) learns compact, yet comprehensive features of pre-detected objects at several levels of abstraction, and performs exhaustive pairing permutations of those features in any two frames to infer object affinities. DAN also accounts for multiple objects appearing and disappearing between video frames. We exploit the resulting efficient affinity computations to associate objects in the current frame deep into the previous frames for reliable on-line tracking. Our technique is evaluated on popular multiple object tracking challenges MOT15, MOT17 and UA-DETRAC. Comprehensive benchmarking under twelve evaluation metrics demonstrates that our approach is among the best performing techniques on the leader board for these challenges. The open source implementation of our work is available at https://github.com/shijieS/SST.git.

Object detection in road scenes is a task that has recently become popular and it is also an important part of intelligent transportation systems. Due to the different locations of cameras in the road scenes, the size of the traffic objects captured varies greatly, which imposes a burden on the network optimization. In addition, in some dense traffic scenes, the size of the traffic objects captured is extremely small and it is easy to miss detection and to encounter false detection. In this paper, we propose an improved multi-scale YOLOv5s algorithm based on the YOLOv5s algorithm. In detail, we add a detection head for extremely small objects to the original YOLOv5s model, which significantly improves the accuracy in detecting extremely small traffic objects. A content-aware reassembly of features (CARAFE) module is introduced in the feature fusion part to enhance the feature fusion. A new SPD-Conv CNN Module is introduced instead of the original convolutional structure to enhance the overall computational efficiency of the model. Finally, the normalization-based attention module (NAM) is introduced, allowing the model to focus on more useful information during training and significantly improving detection accuracy. The experimental results demonstrate that compared with the original YOLOv5s algorithm, the detection accuracy of the multi-scale YOLOv5s model proposed in this paper is improved by 7.1% on the constructed diverse traffic scene datasets. The improved multi-scale YOLOv5s algorithm also maintains the highest detection accuracy among the current mainstream object detection algorithms and is superior in accomplishing the task of detecting traffic objects in complex road scenes.

Vision-based automatic counting of people has widespread applications in intelligent transportation systems, security, and logistics. However, there is currently no large-scale public dataset for benchmarking approaches on this problem. This paper fills this gap by introducing the first real-world RGB-D people counting dataset (PCDS) containing over 4500 videos recorded at the entrance doors of buses in normal and cluttered conditions. It also proposes an efficient method for counting people in real-world cluttered scenes related to public transportations using depth videos. The proposed method computes a point cloud from the depth video frame and re-projects it onto the ground plane to normalize the depth information. The resulting depth image is analyzed for identifying potential human heads. The human head proposals are meticulously refined using a 3D human model. The proposals in each frame of the continuous video stream are tracked to trace their trajectories. The trajectories are again refined to ascertain reliable counting. People are eventually counted by accumulating the head trajectories leaving the scene. To enable effective head and trajectory identification, we also propose two different compound features. A thorough evaluation on PCDS demonstrates that our technique is able to count people in cluttered scenes with high accuracy at 45 fps on a 1.7-GHz processor, and hence it can be deployed for effective real-time people counting for intelligent transportation systems.

Students’ performance prediction is a crucial task in today's online education. By predicting a student's final grade in an academic examination, intervene can be applied in advance. Recently, many machine learning models have been designed to couple students’ online activity with their academic performance. However, it is difficult for these models to effectively make prediction due to the excessive difference in feature selection. While in most cases, too many parameters and heterogeneous features can also be one of the main sticking points. To this end, we propose a sequential engagement based academic performance prediction network. It consists of two main components: an engagement detector and a sequential predictor. The engagement detector leverages the advantages of a convolutional neural network to detect students’ engagement patterns through their daily activities. The sequential predictor adopts the structure of long short-term memory and learns the interaction from the engagement feature spaces and demographic features. By comparing with various existing advanced machine learning models, the results show that this method has better performance than the existing ones when involving the engagement detection mechanism.