Abstract: A method and apparatus of a device that classifies an image is described. In an exemplary embodiment, the method includes tiling at least one region of interest of the input image into a set of tiles. For each tile, the method includes extracting a feature vector of the tile by applying a convolutional neural network, wherein a feature is a local descriptor of the tile; and computing a score of the tile from the extracted feature vector, said tile score being representative of a contribution of the tile into a classification of the input image. The method also includes sorting a set of the tile scores and selecting a subset of the tile scores based on their value and/or their rank in the sorted set. The method also includes applying a classifier to the selected tile scores in order to classify the input image.

Abstract: A method and apparatus of a device that classifies an image is described. In an exemplary embodiment, the method includes tiling at least one region of interest of the input image into a set of tiles. For each tile, the method includes extracting a feature vector of the tile by applying a convolutional neural network, wherein a feature is a local descriptor of the tile; and computing a score of the tile from the extracted feature vector, said tile score being representative of a contribution of the tile into a classification of the input image. The method also includes sorting a set of the tile scores and selecting a subset of the tile scores based on their value and/or their rank in the sorted set. The method also includes applying a classifier to the selected tile scores in order to classify the input image.

Abstract: A method and apparatus of a device that classifies an image is described. In an exemplary embodiment, the method includes tiling at least one region of interest of the input image into a set of tiles. For each tile, the method includes extracting a feature vector of the tile by applying a convolutional neural network, wherein a feature is a local descriptor of the tile; and computing a score of the tile from the extracted feature vector, said tile score being representative of a contribution of the tile into a classification of the input image. The method also includes sorting a set of the tile scores and selecting a subset of the tile scores based on their value and/or their rank in the sorted set. The method also includes applying a classifier to the selected tile scores in order to classify the input image.

Abstract: A method and apparatus of a device that classifies an image is described. In an exemplary embodiment, the method includes tiling at least one region of interest of the input image into a set of tiles. For each tile, the method includes extracting a feature vector of the tile by applying a convolutional neural network, wherein a feature is a local descriptor of the tile; and computing a score of the tile from the extracted feature vector, said tile score being representative of a contribution of the tile into a classification of the input image. The method also includes sorting a set of the tile scores and selecting a subset of the tile scores based on their value and/or their rank in the sorted set.

Abstract: Systems and methods for capturing light field information including spatial and angular information using an image pickup device that includes an image sensor and at least one spatial light modulator (SLM) take multiple captures of a scene using the at least one SLM to obtain coded projections of a light field of the scene, wherein each capture is taken using at least one pattern on the at least one SLM, and recover light field data using a reconstruction process on the obtained coded projections of the light field.

Abstract: Systems and methods for capturing light field information including spatial and angular information using an image pickup device that includes an image sensor and at least one spatial light modulator (SLM) take multiple captures of a scene using the at least one SLM to obtain coded projections of a light field of the scene, wherein each capture is taken using at least one pattern on the at least one SLM, and recover light field data using a reconstruction process on the obtained coded projections of the light field.