Purpose: This study aims to detect research fronts of a scientific field in its early development stage using bibliometric indicators developed based on the features of research fronts.
　　Design/methodology/approach: We first focused on conceptualization of “research fronts” and outlined 4 indicators to measure different aspects of the concept: Innovation, timeliness,stability and interdisciplinarity. Then research specialties obtained based on direct citationnetworks were evaluated with the 4 indicators and research fronts were identified. Finally,methodology was illustrated and verified in the research domain of complex networks.
　　Findings: The results indicate that the proposed method is useful in detecting research fronts in a specific field which is at an early development stage.
　　Research limitations: The quality of clustering results was not assessed. It needs to be further studied whether the method can be applicable to the fields that evolve slowly such as mathematics or are highly specialized like entomology and zoology.
　　Practical implications: This study helps researchers keep abreast of emerging trends and policy-makers formulate science and technology policies and strategic plans.
　　Originality/value: This research provides insights into a quantitative assessment method to detect research fronts of a scientific field in its early development stage.