Search

In this paper we present T1K+, a very large, heterogeneous database of high-quality texture images acquired under variable conditions. T1K+ contains 1129 classes of textures ranging from natural subjects to food, textile samples, construction materials, etc. T1K+ allows the design of experiments especially aimed at understanding the specific issues related to texture classification and retrieval. To help the exploration of the database, all the 1129 classes are hierarchically organized in 5 thematic categories and 266 sub-categories. To complete our study, we present an evaluation of hand-crafted and learned visual descriptors in supervised texture classification tasks.

This paper reviews the first NTIRE challenge on perceptual image enhancement with the focus on proposed solutions and results. The participating teams were solving a real-world photo enhancement problem, where the goal was to map low-quality photos from the iPhone 3GS device to the same photos captured with Canon 70D DSLR camera. The considered problem embraced a number of computer vision subtasks, such as image denoising, image resolution and sharpness enhancement, image color/contrast/exposure adjustment, etc. The target metric used in this challenge combined PSNR and SSIM scores with solutions' perceptual results measured in the user study. The proposed solutions significantly improved baseline results, defining the state-of-the-art for practical image enhancement.

This paper reviews the second NTIRE challenge on image dehazing (restoration of rich details in hazy image) with focus on proposed solutions and results. The training data consists from 55 hazy images (with dense haze generated in an indoor or outdoor environment) and their corresponding ground truth (haze-free) images of the same scene. The dense haze has been produced using a professional haze/fog generator that imitates the real conditions of haze scenes. The evaluation consists from the comparison of the dehazed images with the ground truth images. The dehazing process was learnable through provided pairs of haze-free and hazy train images. There were ~ 270 registered participants and 23 teams competed in the final testing phase. They gauge the state-of-the-art in image dehazing.

We present here a method for computational color constancy in which a deep convolutional neural network is trained to detect achromatic pixels in color images after they have been converted to grayscale. The method does not require any information about the illuminant in the scene and relies on the weak assumption, fulfilled by almost all images available on the web, that training images have been approximately balanced. Because of this requirement we define our method as quasi-unsupervised. After training, unbalanced images can be processed thanks to the preliminary conversion to grayscale of the input to the neural network. The results of an extensive experimentation demonstrate that the proposed method is able to outperform the other unsupervised methods in the state of the art being, at the same time, flexible enough to be supervisedly fine-tuned to reach performance comparable with those of the best supervised methods.

In this work we present SpliNet, a novel CNNbased method that estimates a global color transform for the enhancement of raw images. The method is designed to improve the perceived quality of the images by reproducing the ability of an expert in the field of photo editing. The transformation applied to the input image is found by a convolutional neural network specifically trained for this purpose. More precisely, the network takes as input a raw image and produces as output one set of control points for each of the three color channels. Then, the control points are interpolated with natural cubic splines and the resulting functions are globally applied to the values of the input pixels to produce the output image. Experimental results compare favorably against recent methods in the state of the art on the MIT-Adobe FiveK dataset. Furthermore, we also propose an extension of the SpliNet in which a single neural network is used to model the style of multiple reference retouchers by embedding them into a user space. The style of new users can be reproduced without retraining the network, after a quick modeling stage in which they are positioned in the user space on the basis of their preferences on a very small set of retouched images.

Filtering has been one of the main approaches to texture analysis since early on. Traditionally, the process involved designing the filters essentially by hand based on some prior knowledge (e.g. perceptual models, optimal mathematical properties, etc.) In this work we propose the use of convolutional networks for refactoring traditional, hand-designed filters. Our method consists of initialising the first convolutional layer of the network with some classic banks of filters, training the network on texture images and retrieve the modified filters. Experimenting with five classes of filters and eight datasets of texture images we show that the refactored filters can be conveniently used ‘off-the-shelf’ to achieve better performance than obtained with the original filters, but at the same computational cost.

We propose a novel method based on Convolutional Neural Networks for content-preserving tone adjustment. The method is at the same time fast and accurate since we decouple the inference of the parameters and the color transform: the parameters are inferred from a downsampled version of the input image and the transformation is applied to the full resolution input. The method includes two steps of image enhancement: the first one is a global color transformation, while the second one is a local transformation. Experiments conducted on the DPED - DSLR Photo Enhancement Dataset, that has been used for the NTIRE19 Image Enhancement Challenge, and on the MIT-Adobe FiveK dataset, that is widely used for image enhancement, demonstrate the effectiveness of the proposed method.

In this work we propose a novel CNN-based method for image enhancement that simulates an expert retoucher. The method is fast and accurate at the same time thanks to the decoupling between the inference of the parameters and the color transformation. Specifically, the parameters are inferred from a downsampled version of the raw input image and the transformation is applied to the full resolution input. Different variants of the proposed enhancement method can be generated by varying the parametric functions used as color transformations (i.e. polynomial, piecewise, cosine and radial), and by varying how they are applied (i.e. channelwise or full color). Experimental results show that several variants of the proposed method outperform the state of the art on the MIT-Adobe FiveK dataset

In this paper we propose a strategy for semi-supervised image classification that leverages unsupervised representation learning and co-training. The strategy, that is called CURL from co-training and unsupervised representation learning, iteratively builds two classifiers on two different views of the data. The two views correspond to different representations learned from both labeled and unlabeled data and differ in the fusion scheme used to combine the image features. To assess the performance of our proposal, we conducted several experiments on widely used data sets for scene and object recognition. We considered three scenarios (inductive, transductive and self-taught learning) that differ in the strategy followed to exploit the unlabeled data. As image features we considered a combination of GIST, PHOG, and LBP as well as features extracted from a Convolutional Neural Network. Moreover, two embodiments of CURL are investigated: one using Ensemble Projection as unsupervised representation learning coupled with Logistic Regression, and one based on LapSVM. The results show that CURL clearly outperforms other supervised and semi-supervised learning methods in the state of the art.

We present a method for the automatic restoration of images subjected to the application of photographic filters, such as those made popular by photo-sharing services. The method uses a convolutional neural network (CNN) for the prediction of the coefficients of local polynomial transformations that are applied to the input image. The experiments we conducted on a subset of the Places-205 dataset show that the quality of the restoration performed by our method is clearly superior to that of traditional color balancing and restoration procedures, and to that of recent CNN architectures for image-to-image translation.

Texture classification has a long history in computer vision. In the last decade, the strong affirmation of deep learning techniques in general, and of convolutional neural networks (CNN) in particular, has allowed for a drastic improvement in the accuracy of texture recognition systems. However, their performance may be dampened by the fact that texture images are often characterized by color distributions that are unusual with respect to those seen by the networks during their training. In this paper we will show how suitable color balancing models allow for a significant improvement in the accuracy in recognizing textures for many CNN architectures. The feasibility of our approach is demonstrated by the experimental results obtained on the RawFooT dataset, which includes texture images acquired under several different lighting conditions.

Texture classification algorithms require generalization abilities in order to be reliably used in real world applications. This paper casts this problem in the domain adaptation setting and presents the first study investigating (a) up to which extent this visual recognition problem suffers from this issue, and (b) the effectiveness of existing domain adaptation algorithms in mitigating it. We focus on domain adaptation methods based on shallow classifiers, and test their performance on deep and non deep features. Results obtained on a newly created domain adaptation texture setup show the superiority of deep features compared to other well known approaches, and highlights the importance of factoring in the domain shift when dealing with textures in the wild.

In this paper we propose an approach based on deep Convolutional Neural Networks (CNNs) to recognize artistic photo filters applied to images. A total of 22 types of Instagram-like filters is considered. Different CNN architectures taken from the image recognition literature are compared on a dataset of more than 0.46 M images from the Places-205 dataset. Experimental results show that not only it is possible to reliably determine whether or not one of these filters has been applied, but also which one. Differently from other tasks, where the fine-tuning of a CNN trained on a different problem is usually good enough, here the fine-tuned AlexNet obtains an accuracy of only 67.5%. We show, instead, that an accuracy of about 99.0% can be obtained by training a CNN from scratch for this specific problem.

Recent studies experienced the use of advanced tools for smart aircraft maintenance and inspection. These tools often require the use of computer-vision based technologies to recognize and track a given aircraft mechanical part in order to make it possible to show additional information to a technician on a suitable display. In this paper we propose a visual recognition module of aircraft mechanical parts that has been included in a prototype system designed for the smart maintenance of the Alenia-Aermacchi M346. The evaluation, carried out on real aircrafts, considers different kind of maintenance operations that require the recognition of 20 different mechanical parts. The visual recognition module has been tested under different imaging conditions and varying the scale and the orientation of the parts of interest. The results confirm the feasibility of our proposal also in such a very challenging and realistic condition.

In this paper, we present a three-stage method for the estimation of the color of the illuminant in RAW images. The first stage uses a convolutional neural network that has been specially designed to produce multiple local estimates of the illuminant. The second stage, given the local estimates, determines the number of illuminants in the scene. Finally, local illuminant estimates are refined by non-linear local aggregation, resulting in a global estimate in case of single illuminant. An extensive comparison with both local and global illuminant estimation methods in the state of the art, on standard data sets with single and multiple illuminants, proves the effectiveness of our method.

In this work we have investigated the use of color descriptors to automatically locate the color target in the scene. Three different local descriptors have been tested. These descriptors are then used to return multiple localization hypotheses and a geometrical and appearance validation are introduced to select the most feasible pose. The experimental results on a public dataset of RAW images containing the Macbeth ColorChecker CC target and acquired in uncontrolled environments, showed that all the descriptors considered benefited from the hypothesis validation introduced. © 2011 Springer-Verlag Berlin Heidelberg.

The analysis of color and texture has a long history in image analysis and computer vision. These two properties are often considered as independent, even though they are strongly related in images of natural objects and materials. Correlation between color and texture information is especially relevant in the case of variable illumination, a condition that has a crucial impact on the effectiveness of most visual descriptors. We propose an ensemble of hand-crafted image descriptors designed to capture different aspects of color textures. We show that the use of these descriptors in a multiple classifiers framework makes it possible to achieve a very high classification accuracy in classifying texture images acquired under different lighting conditions. A powerful alternative to hand-crafted descriptors is represented by features obtained with deep learning methods. We also show how the proposed combining strategy hand-crafted and convolutional neural networks features can be used together to further improve the classification accuracy. Experimental results on a food database (raw food texture) demonstrate the effectiveness of the proposed strategy.

The recognition of color texture under varying lighting conditions remains an open issue. Several features have been proposed for this purpose, ranging from traditional statistical descriptors to features extracted with neural networks. Still, it is not completely clear under what circumstances a feature performs better than others. In this paper, we report an extensive comparison of old and new texture features, with and without a color normalization step, with a particular focus on how these features are affected by small and large variations in the lighting conditions. The evaluation is performed on a new texture database, which includes 68 samples of raw food acquired under 46 conditions that present single and combined variations of light color, direction, and intensity. The database allows us to systematically investigate the robustness of texture descriptors across large variations of imaging conditions.

We propose a new self-adaptive image cropping algorithm exploiting both visual and semantic information. Visual information is obtained by a visual attention model, while semantic information relates to the automatically assigned image genre and to the detection of face and skin regions. ©2007 IEEE.

In this work we describe a Convolutional Neural Network (CNN) to accurately predict the scene illumination. Taking image patches as input, the CNN works in the spatial domain without using hand-crafted features that are employed by most previous methods. The network consists of one convolutional layer with max pooling, one fully connected layer and three output nodes. Within the network structure, feature learning and regression are integrated into one optimization process, which leads to a more effective model for estimating scene illumination. This approach achieves state-of-the-art performance on a standard dataset of RAW images. Preliminary experiments on images with spatially varying illumination demonstrate the stability of the local illuminant estimation ability of our CNN.

The description of color texture under varying lighting conditions is still an open issue. We defined a new color texture descriptor, that we called Local Angular Patterns, specially designed to be robust to changes in the color of the illuminant. The results show that our descriptor outperforms the state-of-the-art on a dataset of food textures.

We propose a strategy for land use classification, which exploits multiple kernel learning (MKL) to automatically determine a suitable combination of a set of features without requiring any heuristic knowledge about the classification task. We present a novel procedure that allows MKL to achieve good performance in the case of small training sets. Experimental results on publicly available data sets demonstrate the feasibility of the proposed approach.

Many imaging applications require that images are correctly orientated with respect to their content. In this work we present an algorithm for the automatic detection of the image orientation that relies on the image content as described by Local Binary Patterns (LBP). The detection is efficiently performed by exploiting logistic regression. The proposed algorithm has been extensively evaluated on more than 100,000 images taken from the Scene UNderstanding (SUN) database. The results show that our algorithm outperformed similar approaches in the state of the art, and its accuracy is comparable with that of human observers in detecting the correct orientation of a wide range of image contents.

We present here, an image description approach based on prosemantic features. The images are firstly represented by a set of low-level features related to their structure and color distribution. Those descriptions are fed to a battery of image classifiers trained to evaluate the membership of the images with respect to a set of 14 overlapping classes. Prosemantic features are obtained by packing together the scores. In this paper we will show how prosemantic features outperform traditional low-level features in a variety of tasks. One is content-based retrieval: we included prosemantic features into the framework of the QuickLook2 image retrieval system. Target search experiments show that the use of prosemantic features, combined with the relevance feedback mechanism of QuickLook2, allows for a more successful and quick retrieval of the query images with respect to low-level features. Moreover, we will show the effectiveness of our features for the browsing and visualization of the results obtained from image search engines.

This paper presents a texture descriptor for color texture classification specially designed to be robust against changes in the illumination conditions. The descriptor combines a histogram of local binary patterns (LBPs) with a novel feature measuring the distribution of local color contrast. The proposed descriptor is invariant with respect to rotations and translations of the image plane and with respect to several transformations in the color space. We evaluated the proposed descriptor on the Outex test suite, by measuring the classification accuracy in the case in which training and test images have been acquired under different illuminants. The results obtained show that our descriptor outperforms the original LBP approach and its color variants, even when these are computed after color normalization. Moreover, it also outperforms several other color texture descriptors in the state of the art. © 2014 Optical Society of America.

Recently, new high-level features have been proposed to describe the semantic content of images. These features, that we call supervised, are obtained by exploiting the information provided by an additional set of labeled images. Supervised features were successfully used in the context of image classification and retrieval, where they showed excellent results. In this paper, we will demonstrate that they can be effectively used also for unsupervised image categorization, that is, for grouping semantically similar images. We have experimented different state-of-the-art clustering algorithms on various standard data sets commonly used for supervised image classification evaluations. We have compared the results obtained by using four supervised features (namely, classemes, prosemantic features, object bank, and a feature obtained from a Canonical Correlation Analysis) against those obtained by using low-level features. The results show that supervised features exhibit a remarkable expressiveness which allows to effectively group images into the categories defined by the data sets’ authors