The rapid advancement of artificial intelligence has significantly propelled the widespread application of image recognition technology in fields such as healthcare and security. As a crucial component of Traditional Chinese Medicine (TCM), processed herbal slices hold substantial clinical importance, rendering the enhancement of their detection efficiency and accuracy highly significant. However, traditional image recognition algorithms struggle to achieve satisfactory results due to challenges including intricate backgrounds, high inter-class similarity, and significant intra-class variation among these slices. To address these limitations, this paper proposes YOLOv7FSC, an enhanced object detection algorithm based on the YOLOv7 architecture, specifically designed for processed herbal slice recognition. Tailored to the complexity of herbal slice morphology, our enhancements to YOLOv7 include: Adopting the FReLU activation function to strengthen fine-grained feature extraction of textural details; Incorporating the SENet attention module into the backbone network to optimize feature weight allocation; Replacing conventional upsampling with the CARAFE module to improve feature map resolution and detail reconstruction capabilities. Experimental results demonstrate that YOLOv7FSC significantly outperforms the baseline YOLOv7 model on processed herbal slice recognition, achieving improvements of 7% in , and 3.6% in :0.95, alongside enhanced overall robustness. Furthermore, we develop an interactive user interface using the PyQt5 framework to enable image uploading and recognition result querying. This system provides practical technical support for automated herbal slice detection and demonstrates tangible application value.