Our application AETROS TRAINER gives you a tool to quickly train and test your own digital image recognition expert for all kinds of use cases. You only need to upload your test data (images) and wait until the system finishes the training session. Once trained, you can verify the results and deploy the network in your own infrastructure. Researchers and data scientists can use AETROS TRAINER to build different neural networks using a graphical interface. The application generates basically valid Python code based on Keras.io,

This paper studies the classification of high-dimensional Gaussian signals from low-dimensional noisy, linear measurements. In particular, it provides upper bounds (sufficient conditions) on the number of measurements required to drive the probability of misclassification to zero in the low-noise regime, both for random measurements and designed ones. Such bounds reveal two important operational regimes that are a function of the characteristics of the source: i) when the number of classes is less than or equal to the dimension of the space spanned by signals in each class, reliable classification is possible in the low-noise regime by using a one-vs-all measurement design; ii) when the dimension of the spaces spanned by signals in each class is lower than the number of classes, reliable classification is guaranteed in the low-noise regime by using a simple random measurement design. Simulation results both with synthetic and real data show that our analysis is sharp, in the sense that it is able to gauge the number of measurements required to drive the misclassification probability to zero in the low-noise regime.

What is the right supervisory signal to train visual representations? Current approaches in computer vision use category labels from datasets such as ImageNet to train ConvNets. However, in case of biological agents, visual representation learning does not require millions of semantic labels. We argue that biological agents use physical interactions with the world to learn visual representations unlike current vision systems which just use passive observations (images and videos downloaded from web). For example, babies push objects, poke them, put them in their mouth and throw them to learn representations. Towards this goal, we build one of the first systems on a Baxter platform that pushes, pokes, grasps and observes objects in a tabletop environment. It uses four different types of physical interactions to collect more than 130K datapoints, with each datapoint providing supervision to a shared ConvNet architecture allowing us to learn visual representations. We show the quality of learned representations by observing neuron activations and performing nearest neighbor retrieval on this learned representation. Quantitatively, we evaluate our learned ConvNet on image classification tasks and show improvements compared to learning without external data. Finally, on the task of instance retrieval, our network outperforms the ImageNet network on recall@1 by 3%

Image recognition is the process of identifying and detecting an object or a feature in a digital image or video. This concept is used in many applications like systems for factory automation, toll booth monitoring, and security surveillance. Specific image recognition applications include classifying digits using HOG features and an SVM classifier (Figure 1). Cross correlation can be used for pattern matching and target tracking as shown in Figure 2. An effective approach for image recognition includes using a technical computing environment for data analysis, visualization, and algorithm development.

Baidu, the Chinese internet giant, has created a new AI program to explore the connection between art and music. The Baidu AI Composer creates original music inspired by different pieces of art, evoking the mood of each picture in a musical representation. According to the promotional video released by Baidu, the Baidu AI Composer uses image recognition, connected to the'world's largest neural network', to identify the subject, mood, and even cultural signifiers of a piece of art. These are filtered through a matrix of hundreds of billions of samples and AI training features, using trillions of parameters, to create a complete and original piece of music inspired by the specific piece of art observed. The AI program first identifies elements of the picture – are people represented, or is the focus on nature, or objects, or is it an abstract piece?

Geographic object-based image analysis (GEOBIA) framework has gained increasing interest recently. Following this popular paradigm, we propose a novel multiscale classification approach operating on a hierarchical image representation built from two images at different resolutions. They capture the same scene with different sensors and are naturally fused together through the hierarchical representation, where coarser levels are built from a Low Spatial Resolution (LSR) or Medium Spatial Resolution (MSR) image while finer levels are generated from a High Spatial Resolution (HSR) or Very High Spatial Resolution (VHSR) image. Such a representation allows one to benefit from the context information thanks to the coarser levels, and subregions spatial arrangement information thanks to the finer levels. Two dedicated structured kernels are then used to perform machine learning directly on the constructed hierarchical representation. This strategy overcomes the limits of conventional GEOBIA classification procedures that can handle only one or very few pre-selected scales. Experiments run on an urban classification task show that the proposed approach can highly improve the classification accuracy w.r.t. conventional approaches working on a single scale.

Linear dimensionality reduction techniques are powerful tools for image analysis as they allow the identification of important features in a data set. In particular, nonnegative matrix factorization (NMF) has become very popular as it is able to extract sparse, localized and easily interpretable features by imposing an additive combination of nonnegative basis elements. Nonnegative matrix underapproximation (NMU) is a closely related technique that has the advantage to identify features sequentially. In this paper, we propose a variant of NMU that is particularly well suited for image analysis as it incorporates the spatial information, that is, it takes into account the fact that neighboring pixels are more likely to be contained in the same features, and favors the extraction of localized features by looking for sparse basis elements. We show that our new approach competes favorably with comparable state-of-the-art techniques on synthetic, facial and hyperspectral image data sets.

Researchers have developed a series of AI-based systems that can interpret pathology images and identify the presence and absence of metastatic cancer. The AI systems could lead to new and improved diagnostic methods and treatment. A group of researchers from Beth Israel Deaconess Medical Center (BIDMC) and Harvard Medical School in Boston have teamed up to develop new diagnostic methods based on artificial intelligence (AI). Humayun Irshad, PhD research fellow at Harvard Medical School and one of the lead authors on the research, says that their group is using all kinds of different computational methods to improve diagnostic techniques. "We are developing robust and efficient computational methods to improve diagnostic and prognostic assessment of pathological samples," Irshad says.

Indian Angel Network (IAN), Indian angel investor network, announced undisclosed investment in Gurgaon-based Staqu, an Artificial Intelligence (AI) focused research startup working in automated image understanding technology. The funding will be used to further build and democratize technology and strengthen the team. Staqu, founded in 2015, comprises of researchers and engineers as a part of its core team. Atul Rai, co-founder and CEO said, "We plan to invest this round to expand the computational strength of our VGrep Lab (AI research lab at Staqu) and fuel it with clusters of GPUs and other technical resources. Currently, we are applying our research to solve pressing problems in the e-commerce sector.

In this paper, we tackle the question of discovering an effective set of spatial filters to solve hyperspectral classification problems. Instead of fixing a priori the filters and their parameters using expert knowledge, we let the model find them within random draws in the (possibly infinite) space of possible filters. We define an active set feature learner that includes in the model only features that improve the classifier. To this end, we consider a fast and linear classifier, multiclass logistic classification, and show that with a good representation (the filters discovered), such a simple classifier can reach at least state of the art performances. We apply the proposed active set learner in four hyperspectral image classification problems, including agricultural and urban classification at different resolutions, as well as multimodal data. We also propose a hierarchical setting, which allows to generate more complex banks of features that can better describe the nonlinearities present in the data.