In this paper, we present a steganalytic method to attack the enhanced BPCS steganography (for short EBPCS) proposed by Niimi et al. The key element of the method is analysis of the bitplane-block complexity difference sequences between stego images and cover images based on Hilbert-Huang transform. The Hilbert transform based characteristic vectors are constructed via empirical mode decomposition of the sequences and the support vector machine classifier is used for classification. Experimental results have demonstrated effectiveness of the proposed steganalytic method in spatial domain images and JPEG2000 format images. According to our best knowledge, this method is the first successful attack of EBPCS and can be effectively used in spatial domain and frequency domain (especially DWT/JPEG2000) images.