Offline Writer Identification Using Convolutional Neural Network Activation Features
This work addresses writer identification, a domain-specific problem in document analysis, with incremental improvements on a challenging dataset.
The paper tackled offline writer identification by using activation features from CNNs as local descriptors, achieving about 0.21 absolute improvement in mAP on the bilingual ICDAR dataset compared to state-of-the-art methods.
Convolutional neural networks (CNNs) have recently become the state-of-the-art tool for large-scale image classification. In this work we propose the use of activation features from CNNs as local descriptors for writer identification. A global descriptor is then formed by means of GMM supervector encoding, which is further improved by normalization with the KL-Kernel. We evaluate our method on two publicly available datasets: the ICDAR 2013 benchmark database and the CVL dataset. While we perform comparably to the state of the art on CVL, our proposed method yields about 0.21 absolute improvement in terms of mAP on the challenging bilingual ICDAR dataset.