Picture Book Page-Turning Detection Algorithm Based on Perceptual Hash
<p>Gao Qun</p> · Academic Journal of Computing & Information Science · 2024
Through the analysis of the requirements for page-turning detection algorithms on smart reading devices, a picture book page-turning detection algorithm based on perceptual hashing is designed and efficiently implemented in C language. The principle of the perceptual hashing algorithm is detailed, along with the implementation process and threshold update strategy. Experimental analysis is conducted to evaluate the effectiveness of the algorithm, and key parameters are selected and optimized. The experimental results show that the algorithm performs well both on general-purpose PCs and embedded devices. On a manually annotated test set, the detection rate reaches 94.5%, providing an implementation idea for page-turning detection on devices. Future work will focus on optimizing performance in special scenarios and improving the robustness of the algorithm.