Research on Internal Defect Detection Algorithm of Ultrasonic Image Based on Attention Mechanism

Scanning Acoustic Microscope is a nondestructive and effective technology to detect the internal defects of chips, which is widely used in IC screening. However, at present, the ultrasonic defect detection mainly focuses on traditional visual inspection, which has the disadvantages of low efficiency and high false detection rate. To solve these problems, this paper proposes an ultrasonic defect YOLO detection algorithm by introducing and comparing three common attention modules (SENet, CBAM and ECANet). On the basis of YOLO-v4-tiny model, the attention module is embedded in the feature extraction backbone network to distinguish the importance of different channels in the feature map, strengthen the key features of the target area, and improve the detection accuracy of the network. The experimental results show that the algorithm based on CBAM attention module has better detection accuracy and efficiency.