Standard test image

About: Standard test image is a research topic. Over the lifetime, 5217 publications have been published within this topic receiving 98486 citations.

Phrase-based Image Captioning

Abstract: Generating a novel textual description of an image is an interesting problem that connects computer vision and natural language processing. In this paper, we present a simple model that is able to generate descriptive sentences given a sample image. This model has a strong focus on the syntax of the descriptions. We train a purely bilinear model that learns a metric between an image representation (generated from a previously trained Convolutional Neural Network) and phrases that are used to described them. The system is then able to infer phrases from a given image sample. Based on caption syntax statistics, we propose a simple language model that can produce relevant descriptions for a given test image using the phrases inferred. Our approach, which is considerably simpler than state-of-the-art models, achieves comparable results in two popular datasets for the task: Flickr30k and the recently proposed Microsoft COCO.

The use of fully polarimetric information for the fuzzy neural classification of SAR images

Abstract: Presents a method, based on a fuzzy neural network, that uses fully polarimetric information for terrain and land-use classification of synthetic aperture radar (SAR) image. The proposed approach makes use of statistical properties of polarimetric data, and takes advantage of a fuzzy neural network. A distance measure, based on a complex Wishart distribution, is applied using the fuzzy c-means clustering algorithm, and the clustering result is then incorporated into the neural network. Instead of preselecting the polarization channels to form a feature vector, all elements of the polarimetric covariance matrix serve as the target feature vector as inputs to the neural network. It is thus expected that the neural network will include fully polarimetric backscattering information for image classification. With the generalization, adaptation, and other capabilities of the neural network, information contained in the covariance matrix, such as the amplitude, the phase difference, the degree of polarization, etc., can be fully explored. A test image, acquired by the Jet Propulsion Laboratory Airborne SAR (AIRSAR) system, is used to demonstrate the advantages of the proposed method. It is shown that the proposed approach can greatly enhance the adaptability and the flexibility giving fully polarimetric SAR for terrain cover classification. The integration of fuzzy c-means (FCM) and fast generalization dynamic learning neural network (DLNN) capabilities makes the proposed algorithm an attractive and alternative method for polarimetric SAR classification.

A classifier design for detecting image manipulations

Abstract: In this paper we present a framework for digital image forensics. Based on the assumptions that some processing operations must be done on the image before it is doctored and an expected measurable distortion after processing an image, we design classifiers that discriminates between original and processed images. We propose a novel way of measuring the distortion between two images, one being the original and the other processed. The measurements are used as features in classifier design. Using these classifiers we test whether a suspicious part of a given image has been processed with a particular method or not. Experimental results show that with a high accuracy we are able to tell if some part of an image has undergone a particular or a combination of processing methods.

Dictionary-Based Face Recognition Under Variable Lighting and Pose

Abstract: We present a face recognition algorithm based on simultaneous sparse approximations under varying illumination and pose. A dictionary is learned for each class based on given training examples which minimizes the representation error with a sparseness constraint. A novel test image is projected onto the span of the atoms in each learned dictionary. The resulting residual vectors are then used for classification. To handle variations in lighting conditions and pose, an image relighting technique based on pose-robust albedo estimation is used to generate multiple frontal images of the same person with variable lighting. As a result, the proposed algorithm has the ability to recognize human faces with high accuracy even when only a single or a very few images per person are provided for training. The efficiency of the proposed method is demonstrated using publicly available databases available databases and it is shown that this method is efficient and can perform significantly better than many competitive face recognition algorithms.

Super-resolution of images based on local correlations

Abstract: An adaptive two-step paradigm for the super-resolution of optical images is developed in this paper. The procedure locally projects image samples onto a family of kernels that are learned from image data. First, an unsupervised feature extraction is performed on local neighborhood information from a training image. These features are then used to cluster the neighborhoods into disjoint sets for which an optimal mapping relating homologous neighborhoods across scales can be learned in a supervised manner. A super-resolved image is obtained through the convolution of a low-resolution test image with the established family of kernels. Results demonstrate the effectiveness of the approach.