SOFTWAREThis page contains links to software releases of various projects related to my research.
Wide residual networks
LocNet: Improving Localization Accuracy for Object Detection
Attend Refine Repeat: Active Box Proposal Generation via In-Out Localization
Object detection via a multi-region & semantic segmentation-aware CNN model
Learning to compare image patches via convolutional neural networks
Source code (Torch, C++, Matlab) and trained models are provided that can be used for the fundamental task of comparing image patches and extracting correspondences between images. It implements a general similarity function for image patches that is learnt from scratch (i.e., without resorting to manually-designed features) based on convolutional neural networks. The proposed approach has been shown to provide state-of-the-art performance on the most commonly used benchmark datasets. You can download the library from the project page.
Inference by Learning: Speeding-up Graphical Model Optimization via a Coarse-to-Fine Cascade of Pruning Classifier
We provide source code for a general and versatile framework that significantly speeds-up graphical model optimization while maintaining an excellent solution accuracy. The proposed approach, refereed as Inference by Learning or in short as IbyL, relies on a multi-scale pruning scheme that progressively reduces the solution space by use of a coarse-to-fine cascade of learnt classifiers. We thoroughly experiment with classic computer vision related MRF problems, where our novel framework constantly yields a significant time speed-up (with respect to the most efficient inference methods) and obtains a more accurate solution than directly optimizing the MRF. You can download the code from the project page.
FastPD MRF optimization library
The provided library can be used as a tool for minimizing the energy of a discrete Markov Random Field (MRF). It relies on a state of the art general optimization algorithm, which is much faster than prior art such as conventional graph cut techniques. You can download the library from here.
Deformable registration and motion estimation library
This is a highly customizable software for state-of-the-art deformable image registration using discrete optimization. It also provides an easy-to-use graphical user interface for dense image and volume registration. It can be downloaded from here.
Image completion/inpainting and texture synthesis
You can also check out an independent implementation (by Darren Lafreniere) of our image completion/inpainting algorithm that has been presented in the following paper:
(published in IEEE Transactions on Image Processing)