In the manufacturing of chips, the accurate and effective detection of internal bubble defects of chips is essential to maintain product reliability. In general, the inspection is performed manually by viewing X-ray images, which is time-consuming and less reliable. To solve the above problems, an improved bubble defect detection model YOLO-Xray based on the YOLOv5 algorithm for chip X-ray images is proposed. First, the chip X-ray images are preprocessed by image segmentation to construct the chip X-ray defect dataset, namely, CXray. Then, in the input stage, the K-means++ algorithm is used to re-cluster the CXray dataset to generate the anchors suitable for our dataset. In the backbone network, a micro-scale detection head is added to improve the capabilities for small defect detection. In the neck network, the bi-direction feature fusion idea of BiFPN is used to construct a new feature fusion network based on the improved backbone to fuse the semantic features of different layers. In addition, the Quality Focal Loss function is used to replace the cross-entropy loss function to solve the imbalance of positive and negative samples. The experimental results show that the mean average precision (mAP) of the YOLO-Xray algorithm on the CXray dataset reaches 93.5%, which is 5.1% higher than the original YOLOv5. Meanwhile, the YOLO-Xray algorithm achieves state-of-the-art detection accuracy and speed compared with other mainstream object detection models. This shows the proposed YOLO-Xray algorithm can provide technical support for bubble defect detection in chip X-ray images. The CXray dataset is also open and available at CXray.

In recent years, the continuously increasing number of space objects has caught great attention and triggered the need for space surveillance for the prevention of potential space collisions. Moving objects may appear as streaks in optical images in some situations. However, most of them are usually unknown so that they lack sufficient prior information, which brings great difficulty to detection. Motivated by the fact that streaks have their unique image characteristics, this article presents a novel detection pattern for streak-like targets in single optical images. First, a space target model from point-like target to streak-like target is studied to obtain characteristics. Next, the background removal method used in software named SExtractor is applied to find target candidates. Then, an operator is used to find local maximum points in each target candidate, and the points in the same candidate are clustered according to three domains including space, intensity, and distribution, to identify whether it contains a streak. Finally, the trajectory line of each cluster is extracted. The proposed method was tested on multiple sequences of collected real optical images, which contain targets of different features and found to obtain superior performance in the probability of detection, probability of false alarm, and receiver operating characteristic (ROC), in comparison to baseline methods. In particular, the setting of thresholds is intuitive because of the designed parameters with concrete meaning in images and the normalization process.

The registration of astronomical images, is one of the key technologies to improve the detection accuracy of small and weak targets during astronomical image-based observation. The error of registration has a great influence on the trace association of targets. However, most of the existing methods for point-matching and image transformation lack pertinence for this actual scene. In this study, we propose a registration algorithm based on geometric constraints and homography, for astronomical images. First, the position changes in stars in an image caused by the motion of the platform where the camera had been stationed, were studied, to choose a more targeted registration model, which is based on homography transformation. Next, each image was divided into regions, and the enclosed stable stars were used for the construction of triangles, to reduce the errors from unevenly distributed points and the number of triangles. Then, the triangles in the same region of two images were matched by the geometric constraints of side lengths and a new cumulative confidence matrix. Finally, a strategy of two-stage estimation was applied, to eliminate the influence of false pairs and realize accurate registration. The proposed method was then tested on sequences of real optical images under different imaging conditions and confirmed to have outstanding performance in the dispersion rate of points, the accuracy of matching, and the error of registration, as compared to baseline methods. The mean pixel errors after registration for different sequences are all less than 0.5 when the approximate rotation angle per image is from 0.58 $\times {10}^{-2}$ to 5.89 $\times {10}^{-2}$.

Although space targets have different shapes, sizes and intensities, their distributions share certain commonalities. However, it is difficult to summarize a generalized distribution function for space targets. Moreover, most of the existing methods based on deep learning are not suitable to use directly because of the size of targets and the cost of manual labeling for a full image. In this paper, we proposed a pattern for space target detection based on a convolutional neural network (CNN) to learn essential features of the targets from data. In the processing stage, the background is estimated and removed. Then, image techniques are used to search and process region proposals. Different sizes of region proposals are recognized by a discriminator, which is built upon a small CNN trained with the data of several specific targets. Finally, a non-maximum suppression (NMS) operation is used to remove redundant targets. In the network structure, to further enhance the influence of the effective area, the parameters calculated from the center region of the input are utilized as guidance information and added to the features before the full connection. Moreover, the bias loss is applied to increase the weights of unique features. The experimental results demonstrate the outstanding performance of the proposed method in terms of the number of detected targets, accuracy rate and false alarm rate compared with baseline methods. In particular, the proposed method has a simple network structure and a lower computational cost which can be further promoted and implemented in actual engineering.

The continuous monitoring of vital signs is one of the hottest topics in healthcare. Recent technological advances in sensors, signal processing, and image processing spawned the development of no-contact techniques such as remote photoplethysmography (rPPG). To solve the common problems of rPPG including weak extracted signals, body movements, and generalization with limited data resources, we proposed a dual-path estimation method based on video magnification and deep learning. First, image processes are applied to detect, track, and magnificate facial ROIs automatically. Then, the steady part of the wave of each processed ROI is used for the extraction of features including heart rate, PTT, and features of pulse wave waveform. The blood pressures are estimated from the features via a small CNN. Results comply with the current standard and promise potential clinical applications in the future.