Recently, image representation built upon Convolutional Neural Network (CNN) has been shown to provide effective descriptors for image search, outperforming pre-CNN features as short-vector representations. Yet such models are not compatible with geometry-aware re-ranking methods and still outperformed, on some particular object retrieval benchmarks, by traditional image search systems relying on precise descriptor matching, geometric re-ranking, or query expansion. This work revisits both retrieval stages, namely initial search and re-ranking, by employing the same primitive information derived from the CNN. We build compact feature vectors that encode several image regions without the need to feed multiple inputs to the network. Furthermore, we extend integral images to handle max-pooling on convolutional layer activations, allowing us to efficiently localize matching objects. The resulting bounding box is finally used for image reranking. As a result, this paper significantly improves existing CNN-based recognition pipeline: We report for the first time results competing with traditional methods on the challenging Oxford5k and Paris6k datasets.

International audienceThe recent literature on visual recognition and image classification has been mainly focused on Deep Convolutional Neural Networks (Deep CNN) and their variants, which has resulted in a significant progression of the performance of these algorithms. Building on these recent advances, this paper proposes to explicitly add translation and scale invariance to Deep CNN-based local representations, by introducing a new algorithm for image recognition which is modeling image categories as a collection of automatically discovered distinctive parts. These parts are matched across images while learning their visual model and are finally pooled to provide images signatures. The appearance model of the parts is learnt from the training images to allow the distinction between the categories to be recognized. A key ingredient of the approach is a softassign-like matching algorithm that simultaneously learns the model of each part and automatically assigns image regions to the model's parts. Once the model of the category is trained, it can be used to classify new images by finding image's regions similar to the learned parts and encoding them in a single compact signature. The experimental validation shows that the performance of the proposed approach is better than those of the latest Deep Convolutional Neural Networks approaches, hence providing state-of-the art results on several publicly available datasets

This paper studies the problem of document image classification. More specifically, we address the classification of documents composed of few textual information and complex background (such as identity documents). Unlike most existing systems, the proposed approach simultaneously locates the document and recognizes its class. The latter is defined by the document nature (passport, ID, etc.), emission country, version, and the visible side (main or back). This task is very challenging due to unconstrained capturing conditions, sparse textual information, and varying components that are irrelevant to the classification, e.g. photo, names, address, etc.First, a base of document models is created from reference images. We show that training images are not necessary and only one reference image is enough to create a document model. Then, the query image is matched against all models in the base. Unknown documents are rejected using an estimated quality based on the extracted document. The matching process is optimized to guarantee an execution time independent from the number of document models. Once the document model is found, a more accurate matching is performed to locate the document and facilitate information extraction. Our system is evaluated on several datasets with up to 3042 real documents (representing 64 classes) achieving an accuracy of 96.6%.

To cite this version:Ronan Abstract. This paper studies the classification of identification documents, which is a critical issue in various security contexts. We address this challenge as an application of image classification, a problematic that received a large attention from the scientific community. Several methods are evaluated and we report results allowing a better understanding of the specificity of identification documents. We are especially interested in deep learning approaches, showing good transfer capabilities and high performances.

We address the problem of retrieving all the images containing a specific object in a large image collection, where the input query is given as a set of representative images of the object. This problem is referred to as multiple queries in the literature. For images described with bag-of-visualwords (BOW), one of the best performing approach amounts to simply averaging the query descriptors.This paper 1 introduces an improved fusion of the object description based on the recent concept of generalized maxpooling and memory vectors, which summarizes a set of vectors by a single representative vector. They have the property of reducing the influence of frequent features. Therefore, we propose to build a memory vector for each set of queries and the membership test is performed with each image descriptor from the database, to determine its similarity with the query representative. This new strategy for multiple queries brings a significant improvement for most of the image descriptors we have considered, in particular with Convolutional Neural Networks (CNN) features.