1 January 2025 in International Journal of Intelligent Systems
Accurate segmentation of metastatic lesions is crucial for improving the quality of patient care, particularly in the context of bone scans. However, existing automated methods, which are predominantly data‐driven, exhibit limited performance and lack interpretability. To address these challenges, we propose a novel two‐stage framework that integrates human domain knowledge with data patterns to enhance CNN‐based metastasis lesion segmentation in bone scans. The proposed method comprises two phases: Stage I detects hotspots in bone scans using a CNN‐based model, while Stage II identifies actual metastases by leveraging clinical knowledge of uptake intensity asymmetry. Our approach incorporates a dual‐sampling scheme inspired by diagnostic patterns and an enhanced feature extractor within the hotspot segmentation network, thus augmenting the detection capabilities of traditional data‐driven CNN models. The assessment of symmetrical uptake intensity starts with the symmetry axis of the trunk in the image, followed by a composite similarity measure that considers both geometric symmetry and intensity consistency. Experimental evaluations on 302 clinical cases reveal that our proposed segmentation network improves the Dice similarity coefficient score by 4.34% compared to the baseline method. Furthermore, integrating clinical knowledge significantly reduces false positives, improving the class pixel accuracy score by 2.39% and demonstrating notable adaptability to other segmentation models. Comparative analysis with existing models for metastasis lesion segmentation demonstrates the superior performance of our approach. By incorporating domain knowledge into data patterns, our method enhances automated segmentation performance and bridges the gap between domain expertise and data‐driven methodologies in the automated analysis of low‐resolution bone scans.
1 September 2023 in Journal of Physics: Conference Series
To avoid wasting time and manpower during the wireless charging docking process of logistics robots of different sizes and models in substations, which may affect the operational efficiency of the robots, this paper proposes a wireless charging docking method for multi-sensor integrated logistics robots in substations. First, multiple data signals, such as infrared sensors, pressure sensors, and ultrasonic sensors, are fused within the coarse positioning interval to obtain the length and width data of the robot. The robot is then controlled to travel along the X-axis towards the pre-set displacement sensor while maintaining its position relative to the wireless charging transmitter coil. In the fine positioning interval, the distance between the displacement sensor and the wireless charging transmitter coil is fixed. A coarse-precision positioning transformation method is used to decelerate the robot as it travels along the X-axis towards the pre-set displacement sensor, and the robot is parked at the wireless charging transmitter coil. Additionally, two sets of laser sensors are used to compare the lateral distance, which is half the difference between the charging station width and the robot width. This assists the robot in aligning along the Y-axis with the wireless charging transmitter coil and completing the wireless charging docking process. The experimental results show that the wireless charging system has superior docking time and stability compared to traditional docking methods. It is also compatible with multiple logistics robots of different sizes for wireless charging, with a docking error of less than 0.3%. This effectively improves the charging efficiency of robots in substations.