This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.

Search for Publication

Year(s) from:  to 
Keywords (separated by spaces):

Latent Log-Linear Models for Handwritten Digit Classification

T. Deselaers, T. Gass, G. Heigold and H. Ney
IEEE Transactions on Pattern Analysis and Machine Intelligence
2011, in press


We present latent log-linear models, an extension of log-linear models incorporating latent variables and we propose two applications thereof: log-linear mixture models and image deformation-aware log-linear models. The resulting models are fully discriminative, can be trained efficiently, and the model complexity can be controlled. Log-linear mixture models offer additional flexibility within the log-linear modeling framework. Unlike previous approaches, the image deformation-aware model directly considers image deformations and allows for a discriminative training of the deformation parameters. Both are trained using alternating optimization. For certain variants convergence to a stationary point is guaranteed and in practice even variants without this guarantee converge and find models that perform well. We tune the methods on the USPS dataset and evaluate on the MNIST dataset demonstrating the generalization capabilities of our proposed models. Our models, although using significantly fewer parameters, are able to obtain competitive results with models proposed in the literature.

Download in pdf format
  author = {T. Deselaers and T. Gass and G. Heigold and H. Ney},
  title = {Latent Log-Linear Models for Handwritten Digit Classification },
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year = {2011},
  month = {},
  pages = {},
  volume = {},
  number = {},
  keywords = {Log-linear models, latent variables, conditional random fields, OCR, image classification},
  note = {in press}