Search

Abstract: In a compact descriptor for visual search only a limited number of local features may be included. The estimated probability for correct match between keypoints provides a good criterion for selection of a subset. [Copyright &y& Elsevier]

This work proposes LDVS, a learnable binary local descriptor devised for matching natural images within the MPEG CDVS framework. LDVS descriptors are learned so that they can be sign-quantized and compared using the Hamming distance. The underlying convolutional architecture enjoys a moderate parameters count for operations on mobile devices. Our experiments show that LDVS descriptors perform favorably over comparable learned binary descriptors at patch matching on two different datasets. A complete pair-wise image matching pipeline is then designed around LDVS descriptors, integrating them in the reference CDVS evaluation framework. Experiments show that LDVS descriptors outperform the compressed CDVS SIFT-like descriptors at pair-wise image matching over the challenging CDVS image dataset. [ABSTRACT FROM AUTHOR]

This paper addresses the problem of training a deep neural network for satellite image segmentation so that it can be deployed over images whose statistics differ from those used for training. For example, in postdisaster damage assessment, the tight time constraints make it impractical to train a network from scratch for each image to be segmented. We propose a convolutional encoder–decoder network able to learn visual representations of increasing semantic level as its depth increases, allowing it to generalize over a wider range of satellite images. Then, we propose two additional methods to improve the network performance over each specific image to be segmented. First, we observe that updating the batch normalization layers’ statistics over the target image improves the network performance without human intervention. Second, we show that refining a trained network over a few samples of the image boosts the network performance with minimal human intervention. We evaluate our architecture over three data sets of satellite images, showing the state-of-the-art performance in binary segmentation of previously unseen images and competitive performance with respect to more complex techniques in a multiclass segmentation task. [ABSTRACT FROM AUTHOR]

Abstract Vehicle Make and Model Recognition (VMMR) deals with the problem of classifying vehicles whose appearance may vary significantly when captured from different perspectives. A number of successful approaches to this problem rely on part-based models, requiring however labor-intensive parts annotations. In this work, we address the VMMR problem proposing a deep convolutional architecture built upon multi-scale attention windows. The proposed architecture classifies a vehicle over attention windows which are predicted to minimize the classification error. Through these windows, the visual representations of the most discriminative part of the vehicle are aggregated over different scales which in fact provide more representative features for the classifier. In addition, we define a loss function accounting for the joint classification error across make and model. Besides, a training methodology is devised to stabilize the training process and to impose multi-scale constraints on predicted attention windows. The proposed architecture outperforms state-of-the-art schemes reducing the model classification error over the Stanford dataset by 1.7 % and improving the classification accuracy by 0.2 % and 0.3 % on model and make respectively over the CompCar dataset. Highlights • VMMR can be addressed by formulating a joint loss function. • Predicting attention window scale improves VMMR Performance. • Multi-scale visual representations can increase VMMR accuracy. • Multi-scale patch training enables generating multi-scale attention windows. [ABSTRACT FROM AUTHOR]