The exploration of reservoir regions has a close relationship with the localization of faults. Although faults can be labeled in seismic volumes by experienced interpreters, such manual interpretation is inefficient when dealing with a dramatically growing amount of collected seismic data. To speed up the interpretation efficiency of faults, in this paper, we propose a method that semi-automatically detects fault surfaces using the Hough transform as well as tracking vectors. In the proposed method, we classify seismic sections into reference and predicted ones by borrowing the concept of I- and B-frames in video-coding techniques. For these two types of seismic sections, we introduce different strategies to delineate faults. In reference sections, we first highlight likely fault regions from corresponding coherence maps and apply the Hough transform to extract the features of faults. After removing false features, we optimally connect remaining features under the constraints of coherence maps. Since the accuracy of fault detection in reference sections depends highly on several parameters, to avoid replicating the tweaking of parameters in all seismic sections, we propose tracking detected faults in reference sections through remaining predicted sections, in which faults are labeled based on estimated tracking vectors and geological constraints. To evaluate the performance of the proposed method, we introduce a fault similarity (FauSIM) index that describes the similarity between detected faults and manually picked faults. The FauSIM index based on the Fr\'echet distance focuses on both the local and global comparisons of faults. Experimental results show that the proposed method has the capability to accurately detect faults in seismic sections, and the tracking process improves interpretation efficiency by eliminating tweaked parameters. In addition, comparisons between faults delineated by various methods and faults manually picked show that the FauSIM index is highly correlated with interpreters' subjective perception.