Abstract

We propose a novel object localization methodology withthe purpose of boosting the localization accuracy of state-of-the-art object detection systems. Our model, given asearch region, aims at returning the bounding box of an ob-ject of interest inside this region. To accomplish its goal,it relies on assigning conditional probabilities to each rowand column of this region, where these probabilities provideuseful information regarding the location of the boundariesof the object inside the search region and allow the accu-rate inference of the object bounding box under a simpleprobabilistic framework. For implementing our localization model, we make use ofa convolutional neural network architecture that is properlyadapted for this task, called LocNet. We show experimentally that LocNet achieves a very significant improvement on the mAP for high IoU thresholds on PASCAL VOC2007 test set and that it can be very easily coupled with recent stateof-the-art object detection systems, helping them to boost their performance. Finally, we demonstrate that our detection approach can achieve high detection accuracy even when it is given as input a set of sliding windows, thus proving that it is independent of box proposal methods.