Abstract: Systems and methods provide for feature selection that combines semantic classification and generative filtering with forward selection. Features from an original feature set are divided into feature subsets corresponding to ranked semantic classes. Additionally, low quality features are removed from consideration. Features are selected for a reduced feature set by iteratively processing the feature subsets using forward selection in an order corresponding to the ranking of the semantic classes. The reduced feature set is used to generate a predictive model.

Abstract: Certain embodiments involve augmenting project data with searchable metadata for facilitating project queries. A method includes receiving metadata of the set of projects and identifying a filter within the metadata. The method also includes generating a first vector representing a first project of the set of projects and generating a second vector representing a second project of the set of projects. Further, the method includes grouping the first vector and the second vector into a project cluster based on the first vector and the second vector being within a threshold distance. The project cluster represents a set of filters associated with a subset of projects. Additionally, the method includes assigning a searchable tag to the project cluster based on the filter being a majority filter of the project cluster. The searchable tag includes metadata that facilitates locating the projects responsive to a query to the set of projects.

Abstract: Certain embodiments involve augmenting project data with searchable metadata for facilitating project queries. A method includes receiving metadata of the set of projects and identifying a filter within the metadata. The method also includes generating a first vector representing a first project of the set of projects and generating a second vector representing a second project of the set of projects. Further, the method includes grouping the first vector and the second vector into a project cluster based on the first vector and the second vector being within a threshold distance. The project cluster represents a set of filters associated with a subset of projects. Additionally, the method includes assigning a searchable tag to the project cluster based on the filter being a majority filter of the project cluster. The searchable tag includes metadata that facilitates locating the projects responsive to a query to the set of projects.

Abstract: Systems and methods provide for classification and ranking of features for a hierarchical dataset. A hierarchical schema of features from the dataset is accessed. A hierarchical rank is assigned to each feature based on its schema level in the hierarchical schema. Additionally, a semantic rank is assigned to each feature using a semantic model having ranked semantic contexts. The semantic rank of a feature is assigned by identifying a semantic context of the feature and assigning the rank of the semantic context as the semantic rank of the feature. A rank is computed for each feature as a function of its hierarchical rank and semantic rank.

Abstract: Systems and methods provide for classification and ranking of features for a hierarchical dataset. A hierarchical schema of features from the dataset is accessed. A hierarchical rank is assigned to each feature based on its schema level in the hierarchical schema. Additionally, a semantic rank is assigned to each feature using a semantic model having ranked semantic contexts. The semantic rank of a feature is assigned by identifying a semantic context of the feature and assigning the rank of the semantic context as the semantic rank of the feature. A rank is computed for each feature as a function of its hierarchical rank and semantic rank.

Abstract: A computer implemented method and apparatus for quantifying color perception. The method comprises generating a color lattice, wherein the color lattice represents a mapping of a non-perceptual color space to a perceptual color space; accessing an image, wherein the image comprises a plurality of pixels; and generating a perceptual DNA for the image based on the color lattice.

Abstract: Systems and methods provide for feature selection that combines semantic classification and generative filtering with forward selection. Features from an original feature set are divided into feature subsets corresponding to ranked semantic classes. Additionally, low quality features are removed from consideration. Features are selected for a reduced feature set by iteratively processing the feature subsets using forward selection in an order corresponding to the ranking of the semantic classes. The reduced feature set is used to generate a predictive model.

Abstract: A computer implemented method and apparatus for a memory efficient approach for decoding progressive JPEG images. The method comprises (a) accessing a progressive JPEG image, wherein the progressive JPEG image comprises a plurality of compressed scans and wherein each scan in the plurality of scans comprises a plurality of compressed rows of image data; (b) decompressing not more than half of the compressed rows of each scan in the plurality of scans into a memory buffer; (c) copying a remaining plurality of compressed rows of each scan in the plurality of scans into a plurality of small memory buffers; (d) decoding, in place, the decompressed rows in the memory buffer; and (e) repeating steps (b)-(d) until a predefined number of rows are remaining in the plurality of small memory buffers, at which time the remaining rows are decompressed and decoded.

Abstract: A computer implemented method and apparatus for a memory efficient approach for decoding progressive JPEG images. The method comprises (a) accessing a progressive JPEG image, wherein the progressive JPEG image comprises a plurality of compressed scans and wherein each scan in the plurality of scans comprises a plurality of compressed rows of image data; (b) decompressing not more than half of the compressed rows of each scan in the plurality of scans into a memory buffer; (c) copying a remaining plurality of compressed rows of each scan in the plurality of scans into a plurality of small memory buffers; (d) decoding, in place, the decompressed rows in the memory buffer; and (e) repeating steps (b)—(d) until a predefined number of rows are remaining in the plurality of small memory buffers, at which time the remaining rows are decompressed and decoded.

Abstract: An system is described for managing bicubic downsampling in a constrained environment. The system may be configured to detect a request for a destination image, access a read-copy pattern for the ratio of downsampling, access the source image, and generate the destination image using the source image and the read-copy pattern.

Abstract: A computer implemented method and apparatus for quantifying color perception. The method comprises generating a color lattice, wherein the color lattice represents a mapping of a non-perceptual color space to a perceptual color space; accessing an image, wherein the image comprises a plurality of pixels; and generating a perceptual DNA for the image based on the color lattice.

Abstract: An system is described for managing bicubic downsampling in a constrained environment. The system may be configured to detect a request for a destination image, access a read-copy pattern for the ratio of downsampling, access the source image, and generate the destination image using the source image and the read-copy pattern.