RUC-Tencent at ImageCLEF 2015: Concept Detection, Localization and Sentence Generation
Xirong Li, Qin Jin, Shuai Liao, Junwei Liang, Xixi He, Yujia Huo, Weiyu Lan, Bin Xiao, Yanxiong Lu, Jieping Xu · CLEF (Working Notes) · 2015
In this paper we summarize our experiments in the Image- CLEF 2015 Scalable Concept Image Annotation challenge. The RUC- Tencent team participated in all subtasks: concept detection and local- ization, and image sentence generation. For concept detection, we ex- periments with automated approaches to gather high-quality training examples from the Web, in particular, visual disambiguation by Hierar- chical Semantic Embedding. Per concept, an ensemble of linear SVMs is trained by Negative Bootstrap, with CNN features as image represen- tation. Concept localization is achieved by classifying object proposals generated by Selective Search. For the sentence generation task, we adopt Google's LSTM-RNN model, train it on the MSCOCO dataset, and ne- tune it on the ImageCLEF 2015 development dataset. We further develop a sentence re-ranking strategy based on the concept detection informa- tion from the rst task. Overall, our system is ranked the 3rd for concept detection and localization, and is the best for image sentence generation in both clean and noisy tracks.