21 April 2023 in IEEE Transactions on Instrumentation and Measurement
Inspection of railway defects is crucial for the safe and efficient operation of trains. Recent advancements in convolutional neural networks have led to the development of many effective detection and segmentation algorithms, however, these algorithms often struggle to balance efficiency and precision. In this paper, we present a rendering-based fully convolutional network that generates segmentation results through a coarse-to-fine approach. This allows our framework to make full use of low-level features while minimizing the number of parameters. Additionally, our network generates segmentation results from multiple scales of the feature map and uses residual connections to improve low-level feature detection. To improve training, we propose a novel method that augments the dataset by cutting and pasting images and corresponding ground truth labels horizontally. To better understand the patterns learned by our model, we also generate importance and uncertainty maps to make our model explainable. Our results show that the proposed method outperforms other state-of-the-art image segmentation methods with a higher frame rate and better performance.
13 April 2023 in IEEE Sensors Journal
In recent years, Unmanned Aerial Vehicle (UAV) is widely used in vehicle tracking. However, the objects in drone pictures always consist of fewer pixels due to their flying height. The lacking of visual information results in unreliable tracking results. Meanwhile, higher flying heights capture more targets, which makes it difficult to perform real-time inference with existing methods. To solve the above problems, we present a lightweight tracking model which employed behavioural information in the multi-vehicle tracking problem. Besides, BVTracker utilizes visual information as well as behavioural information independently and consists of two branches. One is a trajectory prediction model based on Long Short-Term Memory (LSTM) network and self-attention network. The other one is an association model base on the Hungarian Algorithm. To enable multi-vehicle tracking, the trajectory of each vehicle is predicted by the trajectory prediction model and the prediction results are associated with the detection findings. Moreover, a new dataset collected by UAVs for vehicles on freeways is applied to train and validate BVTracker. Compared with the state-of-art tracking algorithms, the experimental results show that the proposed method is superior to the existing algorithms at different frame rates and can achieve both robustness and real-time requirements, which profits the fast and effective traffic data analysis.
28 August 2024 in IEEE Transactions on Instrumentation and Measurement
There are a huge number of models claim to enhance visual surface defect inspection accuracy. However, as these models generally function directly within the pixel space, optimizing advanced segmentation techniques frequently demands substantial computational resources and poses challenges for inference on devices with limited computing power. Additionally, many current methodologies are deeply reliant on extensive surface defect datasets. In response to these challenges, our research presents a novel approach based on an auto-encoder structure that uses “latent space” to refine defect segmentation. Within the encoder segment of the autoencoder, we’ve incorporated contrastive learning, amplifying both feature extraction and segmentation capabilities. This architectural choice not only tailors the strategy for prompt response scenarios and underscores its precision in high-accuracy applications, but also addresses the challenges posed by the scarcity of defect samples. As a means to assess our approach and better cater to industrial applications that prioritize sample-level accuracy, we introduce innovative sample-level metrics, namely Mostly Segmented (MS) and Mostly Lost (ML). Experiments conducted on the RSDD and Neuseg datasets underscore the strategy’s steadfast performance under diverse data circumstances. Synthesizing the benefits of the latent space and the contrastive learning, this paper delineates proficient methodology for surface defect segmentation.
1 January 2026 in IEEE Transactions on Intelligent Transportation Systems